Update Article: 308

Original Title

Sanitized Title

Clean Title

Source ID

Article Id01

Article Id02

Corpus ID

Dup

Dup ID

Url

Publication Url

Download Url

Original Abstract

In the last 5 years, the availability of large audio datasets in African
countries has opened unlimited opportunities to build machine intelligence (MI)
technologies that are closer to the people and speak, learn, understand, and do
businesses in local languages, including for those who cannot read and write.
Unfortunately, these audio datasets are not fully exploited by current MI
tools, leaving several Africans out of MI business opportunities. Additionally,
many state-of-the-art MI models are not culture-aware, and the ethics of their
adoption indexes are questionable. The lack thereof is a major drawback in many
applications in Africa. This paper summarizes recent developments in machine
intelligence in Africa from a multi-layer multiscale and culture-aware ethics
perspective, showcasing MI use cases in 54 African countries through 400
articles on MI research, industry, government actions, as well as uses in art,
music, the informal economy, and small businesses in Africa. The survey also
opens discussions on the reliability of MI rankings and indexes in the African
continent as well as algorithmic definitions of unclear terms used in MI.Comment: Accepted and to be presented at DSAI 202

Clean Abstract

Tags

Original Full Text

Machine Intelligence in Africa: a surveyAllahsera Auguste Tapo, Ali Traoré, Sidy Danioko, Hamidou Tembine ∗AbstractIn the last 5 years, the availability of large audio datasets in Africancountries has opened unlimited opportunities to build machine intelli-gence (MI) technologies that are closer to the people and speak, learn,understand, and do businesses in local languages, including for those whocannot read and write. Unfortunately, these audio datasets are not fullyexploited by current MI tools, leaving several Africans out of MI busi-ness opportunities. Additionally, many state-of-the-art MI models are notculture-aware, and the ethics of their adoption indexes are questionable.The lack thereof is a major drawback in many applications in Africa. Thispaper summarizes recent developments in machine intelligence in Africafrom a multi-layer multiscale and culture-aware ethics perspective, show-casing MI use cases in 54 African countries through 400 articles on MIresearch, industry, government actions, as well as uses in art, music, theinformal economy, and small businesses in Africa. The survey also opensdiscussions on the reliability of MI rankings and indexes in the Africancontinent as well as algorithmic definitions of unclear terms used in MI.Keywords: machine intelligence, strategy, learning, risk-awareness, machineintelligence integrity, mean-field-type game theory.1 IntroductionMachine intelligence (MI) focuses on the creation of models, evolutionary dy-namics, and algorithms that enable machines or software to co-learn from dataand improve their performance over time [1, 2]. MI is therefore an advanced com-puter & information science that allows a machine, device, software, program,code, or algorithm to interact intelligently with its environment, which meansit can take measures, make decisions, perform actions, and develop strategiesto maximize its chances of successfully achieving its preferences and objectives[3, 4, 5, 6, 7, 8, 9].Demystifying MI for the General PublicAddressing the global gap in public awareness surrounding MI, including itsuses, benefits, risks, and limitations, is crucial not only in Africa but world-∗Corresponding author: H. Tembine, Learning & Game Theory Laboratory, TIMADIE.Email: tembine@landglab.com1arXiv:2402.02218v1  [cs.CY]  3 Feb 2024wide. The confusion arising from associating the term “intelligence” with fewillustrative machine learning applications, influenced by science fiction and MIbusiness narratives, has led to widespread uncertainty and, in some cases, fear.To unlock the potential benefits that MI holds for Africa, it is imperative tocommence an educational journey for the general public, creating a foundationof informed and knowledgeable users of MI systems. This broader understand-ing will, in turn, contribute to nurturing the technical professionals and highlyskilled specialists necessary to propel the countries’ ambitious MI plans forward.Initiating general public awareness efforts must begin at the grassroots levels ofsociety, ensuring that programs and content are accessible even to those withlimited or no formal education. Recognizing the prevalence of informal edu-cation in many African countries, a variety of courses and training programsshould be tailored to assist recipients in ascending the capacity-building pyra-mid. Designing general public awareness programs that are easily consumableis key, primarily through short videos, brief audios in local languages, or inter-active tradition-inspired games. These formats aim to help the audience graspfundamental MI concepts and distinguish between myth and reality.Regrettably, up to now, the educational journey for the general African publicremains inaccessible due to the absence of machine intelligence that supports thelocal language - an audio-rich language in many areas.Here the objective is not to rank countries by their MI strategy. Some verysmall MI projects had very big social impact in the local population and somesignificant MI Innovation had almost no impact for the local population so far.What we learn from basic game theory and Pareto optimality is that when mul-tiple interdependent objectives are involved as it is the case in MI, the scalar-ization technique which maps the vector of objectives into a single number, isnot necessarily a good idea. This can be observed from the fact that the vector(1, 0) is not better than (0, 1) and vice-versa. These vector elements are oftenreplaced by the well-being elements of the local population, which include mon-etary, non-monetary, technological, non-technological, technical, non-technical,empathetic, etc, which are not captured by a single number.Literature ReviewIn [10], the authors examine MI in Africa with a special focus on challenges andopportunities. The developments in MI have the potential to disrupt and trans-form socio-economic activities across industries. The countries in the GlobalSouth such as those in Africa need to tackle governance issues and lack of in-stitutional capacity to establish the building blocks to allow MI to flourish forthem by them. It is important to also examine the roles of international commu-nities in bridging the technological gaps in Africa by adopting a problem-drivenapproach where local needs and problems are contextualized into MI policy for-mulation rather than a blanket copy-and-paste practice that has limited theadvancement of development policies in Africa. A problem-driven approachwould help African countries to formulate robust MI policies that are relevantto their unique circumstances.2The deployment of basic MI technologies is proliferating on the African con-tinent, but policy responses are still at their early stages. The work in [11] pro-vides an overview of the main elements of MI deployment in Africa, MI’s corebenefits and challenges in African settings, and MI’s core policy dimensions forthe continent. The authors argued that for MI to build, rather than undermine,socio-economic inclusion in African settings, policymakers need to be cognisantof the following key dimensions: gender equity, cultural and linguistic diversity,and labour market shifts.The work in [12] proposes a decolonized appropriation of MI in Africa. Tech-nocoloniality occurs when the use of technology reinforces a colonial mindset,aiming to assert power, control, and domination, often replicating historicalpatterns of oppression. Take digital technologies like the internet and mobilephones, for instance, which are deeply intertwined with the legacy of colonial-ism. This integration, originating from the West, has been imposed on otherregions, notably the African continent. An illustration is the mobile app ’FreeBasics’ by Facebook, allowing users limited access to certain websites withoutdata charges. However, many of these accessible sites promote services of pri-vate US companies and are not available in major African languages. This apphas faced significant criticism as a manifestation of digital colonialism.Despite dominant narratives discouraging anthropomorphizing MI, ShokoSuzuki argues against universal models for our attitudes towards MI [13]. Weargue here that the same applies to MI ethics. We should not expect to havea one-size-fits-all in ethics. Within Africa, we should not expect to have aone-size-fits-all ethics either.From a regional perspective, several studies have shown that machine learn-ing technology can help address some of Africa’s most pervasive problems, suchas poverty alleviation, improving education, delivering quality healthcare ser-vices, and addressing sustainability challenges like food security and climatechange. In [14] , a critical bibliometric analysis study is conducted, coupledwith an extensive literature survey on recent developments and associated ap-plications in machine learning research with a perspective on Africa. The pre-sented bibliometric analysis study consists of 2761 machine learning-related doc-uments, of which 89% were articles with at least 482 citations published in 903journals during the past three decades. Furthermore, the collated documentswere retrieved from the Science Citation Index Expanded , comprising researchpublications from 54 African countries between 1993 and 2021. The bibliomet-ric study shows the visualization of the current landscape and future trendsin machine learning research and its application to facilitate future collabora-tive research and knowledge exchange among authors from different researchinstitutions scattered across the African continent.Agriculture is considered as the main source of food, employment and eco-nomic development in most African countries and beyond. In agricultural pro-duction, increasing quality and quantity of yield while reducing operating costsis key. To safeguard sustainability of the agricultural sector in Africa and glob-ally, farmers need to overcome different challenges faced and efficiently use theavailable limited resources. Use of technology has proved to help farmers find3solutions for different challenges and make maximum use of the available lim-ited resources. Blockchains, internet of things and machine learning innovationsare benefiting farmers to overcome different challenges and make good use ofresources. In [15] , a wide-ranging review of recent studies devoted to applica-tions of internet of things and machine learning in agricultural production inAfrica is presented. The studies reviewed focus on precision farming, animaland environmental condition monitoring, pests and crop disease detection andprediction, weather forecasting and classification, and prediction and estimationof soil properties. The work in [16] explores the realization of MI potential inAfrica, emphasizing the pivotal role of trust. The work in [17] discusses thefailure of mass-mediated feminist scholarship in Africa, highlighting normalizedbody-objectification as a consequence of MI. The work in [18] focuses on usingartificial intelligence for diabetic retinopathy screening in Africa. The work in[19] delves into ethical considerations surrounding the implementation of arti-ficial intelligence in Africa’s healthcare. The work in [20] investigates the useand impact of MI on climate change adaptation in Africa. The work in [21]maps policy and capacity for MI development in Africa. The authors in [22]explore the role of information and communication technologies, including MI,in the fight against money laundering in Africa. The work in [23] addresses themalicious use of MI in Sub-Saharan Africa, posing challenges for Pan-Africancybersecurity. The work in [24] conducts a needs assessment survey for artifi-cial intelligence in Africa. The work in [25] discusses the integration of MI inmedical imaging practice from the perspectives of African radiographers. Theauthors in [26] outline an agenda for journalism research on artificial intelligencein Africa. The work in [27] explores the ethics of MI in Africa, focusing on therule of education. The work in [28] examines artificial intelligence policies inAfrica over the next five years. The work in [29] discusses machine ethics andAfrican identities, offering perspectives on MI in Africa. In [30] the authorsadvocates scaling up MI to curb infectious diseases in Africa. The work in [31]addresses emerging challenges of artificial intelligence in Africa, presented in thecontext of responsible MI. The work in [32] scrutinizes China’s role as a ’digitalcolonizer’ in Africa, focusing on MI’s impact.The paper in [33] assesses the effect of information and communication tech-nologies on the informal economy. The authors applied the Generalized Methodof Moments on a sample of 45 African countries from 2000 to 2017. According tothe findings, the use of ICTs (mobile phone and internet) decreases the spreadof the informal economy in Africa. These results are robust to a battery ofrobustness checks. Furthermore, the results of the mediation analysis show thatthe effect of ICTs on the informal economy is mediated by financial develop-ment, human capital and control of corruption. From a policy perspective, theauthors suggested a quantitative and qualitative consolidation of technologicalinfrastructures, for a sustainable mitigation of the rise of the informal sector inAfrica.In the context of the African continent, machine learning has been usedin ecology [38], gold mining [39], education [40], construction [41], electricity[42], mineral resources [43], wheat [44], optical network [45], Teaching [46], fact-4checking [47], news production [48], digital humanism [49], music [50], farming[51], low-resource languages [52], architecture [53], clinical prediction [54], text-based emotion [55], audio-based emotion [56], Computer Vision Community forAfricans and by Africans [57, 58].ContributionThe objective of this article is to present some advances in MI research, MIuse cases in small businesses, and government actions in Africa. In countries,whenever available, we also highlight MI in industry, art, music, and the infor-mal economy. As we will see, each country has its intrinsic path and culturaladaptation to MI and other emerging technologies such as Graphchains and theInternet of People.Table 1: Some countries with a national MI strategy as of December 2023Country Explicit national MI Strategy reportedAlgeria ✓ 2021Senegal ✓ 2023Benin ✓ 2023Rwanda ✓ 2023Mauritius ✓ 2019Nigeria ✓ 2023Egypt ✓ 2021Tunisia ✓ 2022Seychelles ✓ 2019Some key findings are below:• Some MI strategies (see Table 1) are dictated from outside the Africancontinent, and some meetings are organized in very expensive hotels wherethe local population cannot afford a single night in a lifetime. Some reportsand recommendations are made after those meetings. Unfortunately, mostof these efforts have zero impact in the field and are almost useless to thelocal population, as they are not aware of them, except for a few selectedpeople surrounding the funding on MI.• We have carefully reviewed over 400 research articles on the use cases of MIin Africa (see Fig. 1). Most of the applications are at the testing level on asmall dataset, with manipulated targets, and some are clearly oversellingtheir findings. A random forest does not provide a magical output asclaimed by many of these papers. As the MI field has made progress interms of algorithms, designs, robustness, convergence, and learnability,these studies need to be updated to more practical MI algorithms thanthese random searches.• Ranking African countries by a single uniform weighting average index isclearly unethical, as it does not take into account data transparency and5concrete actions taken by governments beyond the speeches at expensivehotels and palaces.Below we list some schools, research institutes, centers and journals on MIestablished within the continent.• Benin: Atlantics AI Labs: Artificial Intelligence Research Centre• Burkina Faso: Interdisciplinary Center of Excellence in AI for Develop-ment• The Artificial Intelligence & Robotics center of excellence (AI&R CoEs)in Addis Ababa Science and Technology University, in Ethiopia.• The Ethiopian Artificial Intelligence Institute (EAII) which is now theAfrican Artificial Intelligence Center of Excellence.• Ghana: National Artificial Intelligence Center• Ghana: Responsible Artificial Intelligence Lab• Ivory Coast: AI and Robotics Center in Yamoussoukro• Nigeria: National Centre for Artificial Intelligence and Robotics• AIISA: Artificial Intelligence Institute of South Africa• Algeria: National School for Artificial Intelligence• Algeria: House of Artificial Intelligence• Egypt: Faculty of Artificial Intelligence• Egyptian Journal of Artificial Intelligence• Moroccan International Center for Artificial Intelligence• Congo: African Research Centre on Artificial Intelligence• Rwanda: Africa’s Centre of Excellence in Artificial Intelligence• Uganda: Artificial Intelligence and Data Science Lab• Artificial Intelligence Centre of Excellence Africa in Kenya• Malawi : Centre for Artificial Intelligence and STEAM - Science, Tech-nology, Engineering, Arts and Mathematics-.We now list some MI companies in Africa:African Foods Nutrition, DataBusiness-AI, CyberLabs Tech, Kumakan ,SOSEB, Saintypay, Kalabaash, Dunia , Qotto, Toto Riibo, Futurafric AI ,DatawareTech, Khalmax Robotics, mNotify, GreenMatics, DigiExt, CYST, CRI,6Figure 1: Keywords of some MI research topics in Africa7Huggle.care, QualiTrace, Guinaga, Grabal, Timadie, SK1 ART, WETEWomen-in-Drones, Tuteria, Kudi AI , Curacel , Codar Tech Africa, Afrikamart , Neural-Sight, AIfluence, Amini, Halkin, Freshee , M-Shule, AI Connect, Qubitica, CashRadix, AgCelerant, Arie Finance, 4Sight, Agrix Tech, Comparoshop, eFarm,KMER MI, DASTUDY, Teranga Capital, Lengo AI, Semoa, Eazy Chain, So-cialGIS, Dobbee Pay, Solimi Fintech , Artybe, Genoskul, WenakLabs, DaTchad,ZereSoft, KivuGreen, BasaliTech, Plano-OneTree, eFarmers, LibraChat, Mpha-lane, Keti, Nalane, Qoloqo, Tincup, Loop, Hyperlink, AfriFeel, Lelapa AI,Lesan, Xineoh, Clevva, Aerobotics, The Gearsh, Credo, Akiba Digital, Bridge-ment, AfricAi , Neurozone, DataProphet, Vulavula, AkilliCon, Dalil, Transfor-maTek, NileCode, FastAutomate, Synapse Analytics, Intixel, DevisionX, DilenyTechnology, MerQ, WideBot, Aphrie, PasHakeem, MoroccoAI, Annarabic, Sigma.AI,AgriEdge, ATLAN, SudanAI, KatYos, WARM, ScorSell, LWATN, AD’VANTAGE,Business & AI, Deepera.AI, AQUA SAFI, Tabiri Analytics, Congretype, Open-banking, Chil AI Lab, Global Auto Systems, Wekebere, Diagnosify, Xpendly,Kwanso, Tribal Credit, Convertedin, DXwand, Sky.Garden, Save-Your-Wardrobe,Wattnow, Ubenwa Health, DataPathology.Each country is doing its own MI path at the research level targeting con-cretes solutions to local problems.Northern Africa:• Algeria: MI applications in AgriTech, water resource management, startupincubators, strategic approach to MI.• Egypt: Diverse MI landscape from automotive cybersecurity to smart agri-culture, tech innovation hubs, advancements in autonomous instrumentsand FastAutomate technologies.• Libya: MI applications in e-commerce, stock market prediction, smartcities, economic and societal aspects.• Morocco: MI in diverse research areas, MI-enabled startups, societal ben-efits from water management to challenges in the insurance sector.• Sudan: MI focus on groundwater quality assessment, disease forecasting.• Tunisia: Strong emphasis on MI in research, applications in textile in-dustry, water management, urban solid waste forecasting, dialectal speechrecognition.Southern Africa:• Botswana: Water billing, diabetic retinopathy screening, solar radiationprediction, gravel loss condition prediction, HIV/AIDS treatment, datamining, ART program success, clinical decision support.• Eswatini: Financial inclusion, diarrhea outbreak prediction, renewableenergy, COVID-19 case prediction, invasive plant study, MI in education,maize crop yield prediction, economic development through technology.8• Lesotho: Land cover mapping, carbon sequestration, soil organic carbonprediction, weather nowcasting, health insurance enrollment, education,electricity demand forecasting, legal frameworks, wage impact, industrymergers, healthcare policy.• Namibia: MI in education, predicting Gender-Based Violence, cybersecu-rity practices in rural areas, discriminating individual animals with tagsin camera trap images, smart irrigation system for efficient farming.• South Africa: Accenture collaboration with Gordon Institute of BusinessScience, MI in Public Sector HR Management, MI-based medical diagno-sis.Middle Africa :• Angola: Seismic inversion, forest fire detection, urban expansion monitor-ing, bioavailable isoscapes.• Cameroon: MI applications in healthcare, agriculture, small businesses,government initiatives.• Central African Republic: IoT-based smart agriculture, MI predictingelectricity mix, machine learning aiding in primate vocalization classifi-cation, smart city development.• Chad: MI for fertility rate forecasting, conflict risk projections, hatespeech detection, entrepreneurial landscape.• Democratic Republic of the Congo: MI in gully erosion assessment, prop-erty tax roll creation, small businesses using MI for climate resilience,forest resource monitoring.• Equatorial Guinea: MI in researching sea level variability, economic di-versification.• Gabon: MI applications in mapping land cover, monitoring coastal ero-sion, forest height estimation.• Republic of the Congo: African Research Centre on MI establishment,focus on research and digital technology.• Sao Tomé and Principe: MI applications in social protection targeting,value chain analysis in agricultureWestern Africa:• Benin: Soil fertility assessment, banana plant disease detection, electricitygeneration forecasting, public health decision-making, suitability mappingfor rice production.9• Burkina Faso: Predicting malaria epidemics, mapping urban development,forecasting energy consumption, exploring mineral resources.• Cabo Verde: Understanding aerosol properties, studying climate changeimpact, estimating salt consumption, monitoring volcanic eruptions.• Côte d’Ivoire: Machine learning for cocoa farmers, progress towards on-chocerciasis elimination.• Gambia: Machine learning models for pneumonia-related child mortality,smart rural water distribution systems.• Ghana: Urban growth assessment, vehicle ownership modeling, sentimentanalysis, blood demand forecasting, internet data usage analysis, sever-ity prediction of motorcycle crashes, effects of artisanal mining, customsrevenue modeling.• Guinea: Predicting viral load suppression among HIV patients, prognosismodels for Ebola patients.• Guinea-Bissau: Biomass relationships, cashew orchard mapping, learningand innovation in smallholder agriculture, automatic speaker recognition.• Liberia: Cloud computing and machine learning for land cover mapping,predicting local violence, scalable approaches for rural school detection.• Mali: Groundwater potential mapping, cropland abandonment analysis,MI in addressing global health challenges, improved recurrent neural net-works for pathogen recognition, market liberalization policy analysis.• Mauritania: MI-driven insights into English studies, desert locust breedingarea identification, business intelligence models for e-Government, remotemonitoring of water points.• Niger: Electrical charge modeling, land use mapping using satellite timeseries, adult literacy and cooperative training program analysis.• Nigeria: Extensive MI applications including diabetes prevalence detec-tion, crude oil production modeling, flood area prediction, food insecu-rity prediction, entrepreneurial success prediction, mobile forensics forcybercrime detection, genre analysis of Nigerian music, terrorism activ-ity prediction, stock market forecasting, poverty prediction using satelliteimagery.• Senegal: Crop yield prediction, resilient agriculture, machine learningfor rice detection, monitoring artisanal fisheries, predicting road accidentseverity, estimating electrification rates, analyzing the energy - climate -economy - population nexus.• Sierra Leone: Initiative for rapid school mapping using MI and satelliteimagery.10• Togo: Novissi program expansion, machine ethics, wind potential evalua-tion, maize price prediction, solar energy harvesting assessment, land usedynamics forecasting, solar energy harvesting evaluation.Eastern Africa:• Burundi: MI research on malaria case prediction, automated image recog-nition for banana plant diseases, industry focus on optimizing LPG usage.• Comoros: MI research in Education.• Djibouti: Research on sky temperature forecasting, deep learning forfracture-fault detection, industry applications in LPG challenges, air travelexperiences.• Eritrea: Predictive lithologic mapping using remote sensing data.• Ethiopia: Machine learning to predict drought, interpretable models forevaporation in reservoirs.• Kenya: MI Made in Africa supporting startups.• Mauritius: MI in education, maritime IoT potential, strategic approachto MI with national strategy, Mauritius MI Council, MI Academy.• Mozambique: Research areas including assessing OpenStreetMap quality,mapping land use and cover, food security, smallholder irrigated agri-culture mapping, deep learning and Twitter for mapping built-up areaspost-natural disasters.• Rwanda: National MI Policy approval.• Tanzania: MI applications in healthcare, Resilience Academy studentsusing machine learning for tree-cover mapping.• Uganda: Creating high-quality datasets for East African languages.• Somalia: Sentiment analysis applied to Somali text.• South Sudan: Machine learning to analyze fragility-related data.• Zambia: Machine learning in predicting stunting among children, enhanc-ing health clinic verification efficiency.• Zimbabwe: Data-driven pediatrics, vehicle damage classification usingdeep learning algorithms, technology to predict and address adolescentdepression.Figures 1 and 1 display the number of occurrences vs MI-related topics andtheir applications in Africa that appears in the title of the 400 references usedin this survey.Examining closely 400 articles on MI in Africa, beyond the headlines, itemerges that11Figure 2: Keywords of some MI research topics in AfricaFigure 3: Application of MI in Africa12• human learning, the learning of men and women, whether children oradults, takes a much more central place in discussions than machine learn-ing.• At the core is human learning, utilizing various tools, including machineassistance, as well as inspiration from nature.Content of the articleThe rest of article is structured as follows. Section 2 is dedicated to NorthernAfrica. Section 3 presents it in Southern Africa. Section 4 presents it in CentralAfrica. Section 5 presents MI research and use cases in Western Africa. Section6 is dedicated to Eastern Africa. Section 7 examines MI indexes and limitationof MI adoptions in Africa. Section 8 discusses technical and non-technical ethicsof MI in Africa. Multi-scale multi-layer multi-modal culture-aware ethics arepresented in Section 9. Section 10 presents data issues and some terms borrowedfrom psychology that are not undefined in the context of computer science.Section 11 concludes the article.2 MI in North AfricaThis section presents some of the remarkable advancements and strategic ini-tiatives undertaken by Algeria, Egypt, Libya, Morocco, Sudan, and Tunisia inharnessing the power of MI. In recent years, these nations have emerged as keyplayers in the ever-evolving realm of MI, propelling the region toward technolog-ical innovation and sustainable growth. From the economic intelligence mech-anisms supporting small and medium enterprises in Algeria to the pioneeringNational MI strategy of Tunisia, each country contributes a unique perspectiveto the broader narrative of MI adoption. We observe diverse facets of MI im-plementation, from educational and healthcare strategies to water managementsolutions and economic reforms. Noteworthy research endeavors, such as model-ing groundwater quality in Sudan and forecasting urban solid waste in Tunisia,showcase the region’s commitment to addressing complex challenges throughcutting-edge technologies. As we navigate through the state of the art in MIacross North Africa, we gain insights into the pivotal role MI plays in shapingthe future of these nations.Algeria exhibits a growing focus on MI applications in AgriTech, emphasizingwater resource management and the emergence of startup incubators. Witha particular emphasis on groundwater quality and smart agriculture, Algeria’sMI initiatives aim to address environmental challenges and enhance agriculturalpractices. The country’s key achievements include the development of AgriTechsolutions and the integration of MI for efficient water resource management.Egypt stands out with a diverse MI landscape, ranging from automotivecybersecurity to smart agriculture. The country’s tech innovation hubs fosterresearch in autonomous instruments and FastAutomate technologies. In theautomotive sector, the focus on cybersecurity aligns with global trends, while13the application of MI in smart agriculture indicates a commitment to lever-aging technology for sustainable practices. Egypt’s MI achievements includeadvancements in autonomous instruments and a robust presence in the realmof FastAutomate technologies.Libya’s MI research landscape spans diverse domains, including the adop-tion of e-commerce in SMEs, stock market prediction, and the enhancementof quality of life through smart cities. Notable studies explore MI applicationsin Libyan SMEs, predicting daily stock market movements with high accuracy,and leveraging MI for smart city development. Libya’s MI initiatives showcasea comprehensive approach, addressing economic, financial, and societal aspectswith the aim of fostering technological integration and improving living stan-dards.Morocco demonstrates a multifaceted approach to MI, covering diverse re-search areas such as data use challenges, MI impact on human rights, and appli-cations like automating water meter data collection. Notable achievements in-clude the development of MI-enabled startups like Annarabic and SYGMA.AI,addressing challenges in insurance claims settlements and revolutionizing theinsurance industry. Morocco’s research efforts highlight a commitment to uti-lizing MI for societal benefits, from enhancing water management to addressingchallenges in the insurance sector.Sudan’s MI focus is evident in groundwater quality assessment and diseaseforecasting. Research endeavors employ MI algorithms like multilayer percep-tron neural networks and support vector regression to evaluate groundwatersuitability for drinking. Additionally, the application of time series forecastingmethodologies aids in predicting diseases like malaria and pneumonia. Sudan’sMI efforts underscore a commitment to addressing critical issues in public healthand environmental sustainability.Tunisia showcases a strong emphasis on MI in research, spanning applica-tions in the textile industry, water management, urban solid waste forecasting,and dialectal speech recognition. The country’s acquisition of Instadeep, a deeptech startup, signifies global recognition of Tunisian innovations. Governmentinitiatives include the integration of MI in public finance management, aligningwith broader strategies for economic growth and accountability. Tunisia’s MIlandscape reflects a dynamic and innovative approach, positioning itself as ahub for entrepreneurship and technological advancements in the North Africanregion.Throughout the period of 2000-2023, Northern African countries have demon-strated a growing embrace of MI technologies, leveraging them across diversesectors to address challenges and foster sustainable development in their uniquesocio-economic contexts.2.1 AlgeriaAlgeria has made significant strides in deploying MI across various sectors from2000 to 2023. Research initiatives highlight the crucial role of economic intel-ligence in supporting the growth of small and medium enterprises , emphasiz-14ing the synergy between economic intelligence and MI for SME development.Algiers serves as a hub for diverse MI-enabled platforms contributing to therise of small and medium businesses, ranging from payroll and HR manage-ment (RAWATIB) to innovative solutions for the visually impaired (Dalil). Inthe informal economy, leveraging MI technologies is proposed to redefine taxpolicies, assess incentives for agriculture, and enhance electronic transactions,contributing to a regulated formal economy. At the governmental level, Algeriahas adopted a national strategy on research and innovation in MI, inauguratedthe National School for Artificial Intelligence (ENSIA), and declared 2023 as’The Year of Artificial Intelligence’, showcasing a comprehensive approach toadvancing skills and leveraging MI in key socio-economic sectors.Algeria 2020-2023 Concrete ActionsResearch ✓ ENSIA, RAWATIBSMB ✓ AkilliCon, Dalil,TransformaTekInformalEconomy✓ Data 1980-2017Government✓ 2021: 100-page whitepaper on national MIstrategyTable 2: MI in Algeria2.1.1 ResearchThe study in [59] underscores the pivotal role of economic intelligence as a cru-cial mechanism in fostering the growth and development of small and mediumenterprises in Algeria. For these enterprises, success hinges on their ability toaccess high-quality and timely information, enabling them to navigate a rapidlychanging environment, anticipate shifts, and make optimal decisions for theirsurvival. The study argues that integrating economic intelligence mechanismsinto SMEs is not just beneficial but an imperative necessity for their develop-ment, support, and enhanced competitiveness. Through a descriptive approachthat emphasizes the need for information transfer, analysis, and drawing fromvarious sources, the results affirm the significant contribution of economic intel-ligence to the competitiveness, innovation, and strategic development of smalland medium enterprises, empowering them to effectively face risks and bolsterdecision-making. In the context of Algeria’s economic landscape, the studyemphasizes the symbiotic relationship between economic intelligence and theMI-driven economy, particularly within the realm of small and medium busi-nesses. The integration of these intelligence mechanisms becomes a strategicimperative, aligning with the broader national strategy for economic develop-ment. As Algeria looks towards the future, the study suggests that the synergybetween economic intelligence and MI holds substantial potential for propelling15the growth and resilience of SMEs, thereby contributing to the overall economicvitality of the country.2.1.2 Small BusinessesAlgiers serves as a focal point for MI-powered platforms like RAWATIB, Akil-liCon, Dalil, and TransformaTek, contributing to the emergence of MI-enabledsmall and medium businesses in Algeria. RAWATIB is an MI-powered SaaSplatform that streamlines payroll and human resources management for busi-nesses of all sizes. The platform automates many administrative tasks involvedin payroll and HR management, reducing the risk of errors and saving busi-nesses time and resources. AkilliCon is a low-cost, low-profile Ambient EnergyHarvester Terminal that can be used alone with AkilliCon specified Battery asa Power Bank. Dalil is a company focused on object recognition and navigationsystems for visually impaired people. Imagine this: your friend invited you todinner at a new restaurant downtown, but to get there, you need to go to thebus stop, take bus number 128 to the train station, and take the train into thecity. Easy enough, right? Now, imagine making that trip without being able tosee. It is a challenge that 314 million visually impaired people face every day.Even with expert mobility skills and the use of a cane or a guiding dog, naviga-tion, environment detection, and recognition can be stressful. TransformaTekis a startup working toward the widespread adoption of location intelligencetechnologies by small businesses. Our mission is to develop community-drivenplatforms to democratize access to open geospatial datasets and build useful usecases for businesses.2.1.3 Informal EconomyThis research work [63] examines Algeria’s informal economy through a com-prehensive analysis employing the Multiple Indicator Multiple Causes approach.The primary objectives include investigating key determinants, estimating sizeand development, examining short and long-term relationships with the formaleconomy, and unraveling causality directions from 1980 to 2017. Findings re-veal that the tax burden, the agricultural sector, the quality of institutions,and GDP per capita emerge as pivotal determinants shaping the contours ofthe IE in Algeria. Understanding these factors becomes imperative for craftingeffective strategies to address the informal sector. The study indicates that theinformal economy constitutes an average of 33.48% of the official GDP, wit-nessing a steady increase over the past 15 years. Interestingly, the researchuncovers a nuanced relationship between the informal and formal economies.In the short run, the IE exerts a positive impact on the formal economy, whilein the long run, this effect undergoes a reversal. Understanding these dynam-ics is crucial for policymakers in devising interventions that promote sustainableeconomic growth.To mitigate the size of the informal economy in Algeria, a mul-tifaceted approach is proposed. Leveraging MI technologies can play a pivotalrole in redefining tax policies, reassessing incentives for the agricultural sector,16and fostering the widespread adoption of electronic transactions. MI-poweredsolutions can enhance efficiency, transparency, and compliance in economic ac-tivities, contributing to a more regulated and inclusive formal economy.2.1.4 GovernmentIn 2019, Algeria organized a workshop on MI. Participants in a workshop forthe preparation of the national strategic plan for MI 2020-2030 recommended,in Constantine, the development of a white paper on this technology to estab-lish economic intelligence in the country. The creation of a white paper on MIserves as a roadmap to determine the appropriate mechanisms for the introduc-tion of this technology into various socio-economic sectors, aiming to achieveeconomic intelligence, according to the consensus of the 150+ Algerian MI ex-perts, including 30 researchers working abroad, gathered at the National Schoolof Biotechnology at Salah Boubnider Constantine University. Subsequently, a100-page white paper was presented to the public in 2021. Algeria has adopteda national strategy on research and innovation in MI, dedicated to improvingAlgerians’ skills in MI through education, training, and research, and exploitingthe potential of MI as a development tool in key socio-economic sectors (e.g.education, health, transport, energy). In the 2021-2022 academic year, Algeriainaugurated the National School for Artificial Intelligence (ENSIA) [62] to admithigh school graduates interested in this field. The school is tailored to educateengineers in the theories and applications of MI and data science. Students willlearn to develop and publish practical and innovative solutions for challengesin sectors such as health, energy, agriculture, and transportation, thereby con-tributing to the country’s scientific and economic advancement. The House ofArtificial Intelligence [61] comprises 13 universities nationwide and supervises40 MI research projects. The Ministry of Higher Education and Scientific Re-search, announced 2023 as ”The Year of MI”, on January 10th at The NationalSchool of Artificial Intelligence (ENSIA).2.2 EgyptEgypt has undergone a transformative journey in deploying MI from 2000 to2023, marked by the innovative endeavors of companies like NileCode, AT-Instruments, and FastAutomate. NileCode positions itself as a comprehensivetech solutions provider, committed to delivering lasting results. In the contextof automotive cybersecurity, AT-Instruments introduces an anomaly detectionsystem with MI algorithms, addressing threats in the expanding attack sur-face. FastAutomate revolutionizes SMB hiring with ”digimates,” MI candidatestrained through job shadowing. Agriculture sees a paradigm shift through the”Hudhud” smart assistant, utilizing MI for disease management, land identi-fication, and insights into informal markets. Government initiatives, such asthe National Council for MI and the National MI Strategy, underscore Egypt’sstrategic focus on MI for sustainable development. Collaborations between theMinistry of Tourism and Antiquities and the Atomic Energy Authority highlight17the pioneering use of MI in cultural preservation, reflecting Egypt’s commitmentto technological innovation across diverse sectors.Egypt 2020-2023 Concrete ActionsResearch ✓ E.J. AISMB ✓ NileCode, FastAu-tomate, SynapseAnalytics, Intixel,DevisionX, DilenyTechnology, MerQ,WideBot, Aphrie,PasHakeemInformalEconomy✓ Tourism, HudhudGovernment✓ 2019: National MIstrategyTable 3: MI in Egypt2.2.1 ResearchThe Egyptian Journal of Artificial Intelligence (E.J.AI) is a biannual refereedjournal issued by the Faculty of Artificial Intelligence - Kafrelsheikh University,which publishes original and state of the art research and developments in thefield of MI and related sciences. Areas of interest may include, but not limitedto, machine learning, deep learning, information retrieval, computer vision, in-telligent machines, robotics, Internet of Things, wireless sensor networks, cloudcomputing, computer science, hardware implementation, image processing andvideo processing (see [65]). In the rapidly evolving landscape of education,the integration of MI has emerged as a transformative force. The scientific re-search work [66] presents MI applications in educational settings, specificallyfocusing on the acceptance and adoption of these cutting-edge tools. As class-rooms evolve into technologically advanced hubs of learning, understanding thedynamics that influence educators and students’ willingness to embrace MI be-comes imperative. The study employs rigorous methodologies, including theUnified Theory of Acceptance and Use of Technology, to navigate the interplaybetween factors such as performance expectancy, effort expectancy, and socialinfluence. By shedding light on the nuanced relationship between MI technol-ogy and its users in education, their research aims to pave the way for informedstrategies that harness the full potential of MI in shaping the future of learningin Egypt. The work in [66] investigates the acceptance of applying chat-bottechnology and related MI technologies, among higher education students inEgypt. Chat-bot, as an MI technology, has garnered significant attention, es-pecially in the education sector. Before implementing such new technology, it18is vital to understand the determinants that affect students’ behavioral inten-tions to accept or reject this technology in higher education. To comprehendthis behavioral intention, the current research applied the Unified Theory ofAcceptance and Use of Technology, excluding two moderators from the origi-nal model - experience and voluntariness of use. Additionally, their researchwork excluded facilitating conditions and behavior use, focusing solely on theintention behavior of students. The research study also examined the role of de-mographic factors (gender and age) in influencing the independent variables ofthe research model and the behavioral intention variable. The research outlinedthe study’s objectives, including developing a framework for the acceptance ofchat-bot technology on the behavioral intention of students in higher educationin Egypt. To achieve these goals, the researcher collected data on the requiredvariables through a questionnaire targeting students at the Arab Academy forScience and Technology and Maritime Transport (AASTMT). AASTMT waschosen due to being one of the oldest private universities in Egypt that imple-ments MI technology in its educational system. The final sample comprised 385responses, and data analysis involved testing, descriptive analysis, correlations,regression, and structural equation modeling. Their Results indicated a signifi-cant impact of performance expectancy, effort expectancy, and social influenceon students’ behavioral intention to accept chat-bot technology in their highereducation in Egypt. Moreover, the results revealed no moderating role of de-mographic factors (gender and age) in the relationship between performanceexpectancy, effort expectancy, social influence, and behavioral intention. Thestudy [67] explores the impact of MI on the tourism industry in Egypt. MI tools,such as chatbots and personalized service recommendations, play a significantrole in travel agencies, affecting various sectors within the tourism industry.The research aims to investigate the implementation of MI techniques in Egyp-tian tourism companies and assess employees’ perceptions of using MI toolsin tourism operations. Utilizing a quantitative approach, the researchers dis-tributed an online questionnaire to tourism companies, with 320 valid responsessubjected to statistical analysis. The results highlight notable differences amongtourism companies offering full services when it comes to applying MI tools intheir operations. Additionally, the size of tourism companies plays a role, withlarger and medium-scale enterprises employing MI techniques more than theirsmaller counterparts. The study identifies two primary employee perspectiveson MI: enthusiasm for its advantages and suspicion regarding its disadvantages.From a managerial perspective, the research sheds light on applied MI tech-niques in tourism and underscores the importance of implementation. Thisinsight can assist managers in formulating policies and strategies to enhancetechnological infrastructure, skills, and the application of beneficial MI tools,ultimately improving performance and saving both time and money.2.2.2 Small BusinessesIn the dynamic landscape of small and medium businesses in Egypt, NileCodeemerges as a seasoned Technology House, boasting a remarkable 15-year journey19in crafting seamless digital experiences. Their extensive team of tech experts,well-versed in diverse fields and grounded in technology consulting, positionsNileCode as a comprehensive 360° provider of authentic tech solutions. Unliketemporary fixes and short-lived prototypes, NileCode is committed to deliv-ering tangible and lasting results for their clients. Drawing inspiration fromthe iconic river Nile, NileCode is on a mission to cultivate rich and flowingexperiences that precisely meet the needs of their business partners. With ablend of innovation, agility, and adaptability, the NileCode team stands outfor its creative prowess and dedicated focus on addressing the persistent chal-lenges encountered by clients. In the context of automotive cybersecurity whichis very important in these geographical areas, Autonomous Instruments calledAT-Instruments introduces an innovative anomaly detection system that runsMI algorithms on GPU. As the automotive attack surface expands rapidly, AT-instruments provides a cutting-edge solution. Its anomaly detection capabilities,coupled with MI algorithms, enable real-time monitoring and swift identifica-tion of anomalies, empowering businesses to counteract potential threats, evenon the day of the attack. FastAutomate, another standout player, introducesa revolutionary approach to MI in the hiring process for SMBs. With a poolof virtual MI candidates, known as ”digimates,” FastAutomate offers a solutionthat goes beyond conventional hiring practices. These Machine employees pos-sess diverse computer skills and are trained through job shadowing. A singlevideo demonstration equips them to perform tasks in uncertain environmentsfaster and with higher accuracy than their human counterparts. FastAutomate’sinnovative approach streamlines the hiring process for SMBs, providing themwith efficient and skilled Machine employees.Synapse Analytics is a B2B IT company in Egypt, assisting organizations inoptimizing workflows through data, machine learning, and MI models. They of-fer a suite of solutions to drive MI adoption across entire organizations. Intixel,a health tech company in Egypt, utilizes MI for medical image analysis, provid-ing modules like Cardiac MRI Segmentation and Skin Cancer Detection. Theirsolutions empower radiologists with efficient, automated second-eye options.DevisionX in Egypt helps businesses integrate MI and computer vision withoutcoding. Their platform enables organizations to label, train, and deploy customMI-vision applications seamlessly. Dileny Technology, based in Egypt, focuseson futuristic healthcare solutions for Africans, developing MI and medical imag-ing systems to enhance diagnoses and treatment administration. MerQ, a techcompany in Egypt, specializes in advanced communications and customer rela-tionship management systems for financial organizations, deploying MI-poweredinteractive programs. WideBot, a B2B CRM solutions provider, assists orga-nizations in creating personalized customer experiences through MI-poweredchatbots and offers services like data training and optimization. Aphrie, an ITcompany, helps businesses leverage MI for operational efficiency, offering servicessuch as web development, mobile apps, quality control, and cloud operations.PasHakeem, a health tech company in Egypt, leverages MI to provide integratedhealthcare services for Africans, supporting telemedicine and efficient medicalrecord-keeping.20Tribal Credit powers startup growth in emerging markets by providing cor-porate cards and financial solutions. Convertedin is an automation ads platformfor eCommerce and online sellers. DXwand is an AI venture founded by technol-ogy & AI experts that are passionate about building intelligent conversationalAI.Together, these forward-thinking entities exemplify the transformative im-pact of MI on small and medium businesses in Egypt, showcasing how technol-ogy can be harnessed to enhance experiences, address challenges, and propelbusinesses into a future of innovation and efficiency.2.2.3 Informal EconomyIn the heart of Egypt’s agricultural landscape, the ”Hudhud” smart assistantproject is reshaping the way farmers engage with their crops. This Arabic mo-bile application utilizes cutting-edge MI techniques, ushering in a qualitativedevelopment in the agricultural extension system. Tailored to individual farm-ers’ needs, crops, and potential pests, ”Hudhud” delivers instantaneous andaccurate guidance, propelling Egypt towards smart, modern agriculture - a cor-nerstone in the construction of a digital Egypt. A notable feature of ”Hudhud”is its ability to empower farmers in identifying and combating crop infectionsswiftly. By capturing a photo of an infected plant and submitting it through thesmartphone application, farmers tap into the power of MI. The app analyzesthe image, identifies the disease, and provides farmers with precise instructionsto halt the infection, offering a real-time solution to a pressing problem. Be-yond disease management, ”Hudhud” contributes to the agricultural sector’sevolution by leveraging drones images and MI to identify reclaimable lands andrecommend optimal crops for each season and region. This strategic approachenhances productivity and resource allocation, aligning with the broader visionof a digitally transformed Egypt. The initiative extends to the distribution of2 million smart farmer cards, a transformative step towards eliminating admin-istrative corruption and ensuring the equitable distribution of subsidies. Eachcard, personalized with the farmer’s name, identity information, and landhold-ing numbers, contributes to the creation of an extensive database encompassingfarmers and cultivated land. This digital infrastructure not only streamlinesadministrative processes but also paves the way for a more transparent andaccountable agricultural sector. ”Hudhud” stands as a beacon for smart agri-cultural practices, guiding farmers through every stage - from cultivation toharvest. By addressing plant diseases, pests, and offering tailored treatments,this MI-driven initiative is propelling Egypt towards a sustainable and techno-logically advanced agricultural future. Furthermore, MI goes beyond the agri-cultural fields, playing a crucial role in shedding light on the dynamics of Egypt’sinformal agricultural markets. The ”Hudhud” project serves as more than justa smart assistant; it acts as a conduit to comprehend the nuances of the infor-mal market. Through the application of MI techniques, it analyzes patterns,market trends, and farmer interactions. This not only enhances the efficiency ofresource allocation but also provides valuable insights into the informal market,21fostering a deeper understanding of the challenges and opportunities within.As Egypt advances towards a digital agricultural landscape, the role of MI indeciphering the complexities of informal markets becomes increasingly indis-pensable, offering a data-driven foundation for informed decision-making andsustainable growth.2.2.4 GovernmentIn November 2019, the Egyptian government formed the National Council for MIas a partnership between governmental institutions, prominent academics andpractitioners from leading businesses in the field of MI. The National Councilfor MI is chaired by the Minister of Communications and Information Technol-ogy. The Council is in charge of outlining, implementing and governing the MIstrategy in close coordination with the concerned experts and entities. See [64].The National MI Strategy 2020-2025 is a key priority to help Egypt achieve itssustainable development goals. It identifies the country’s plans to deepen theuse of MI technologies and transform the economy. It goes beyond just adopt-ing technology, to fundamentally rethinking business models and making deepchanges to reap the benefits of productivity growth and create new areas ofgrowth. In 2023 In Egypt, the Ministry of Tourism and Antiquities collaborateswith the Atomic Energy Authority to pioneer a workshop utilizing nuclear, ra-diological, and MI techniques for the restoration and documentation of ancientmummies and human remains. The workshop, organized by the Department ofRestoration of Mummies and Human Remains, involves experts from variousinstitutions. The Atomic Energy Authority’s advanced laboratories, equippedwith cutting-edge technologies, allow non-invasive examination and analysis ofhuman remains. The MI Division of the authority supports restorers in theirtasks. This marks the first use of MI technology on mummies in Egypt, aimingnot only for video preparation and facial reconstruction but also for enhancingrestoration processes. The program enables restoration workers to evaluate therestoration process before commencement, using special software on laptops oriPads to photograph and plan the restoration of available bones.2.3 LibyaIn Libya, various research works have explored MI applications across differentdomains. Ahmed Naji & Ahmed Abu Aeshah’s research delves into the adop-tion of e-commerce in small and medium enterprises, emphasizing the benefitsof MI in marketing, finance, data capture, and employee relationships. NajebMasoud’s work focuses on predicting movements in the Libyan Stock Marketusing a machine neural network model, showcasing its accuracy in forecastingdaily stock market prices. Ali AA Alarbi, Dani Strickland, and Richard Blan-chard explore MI-enabled demand side management to intelligently address loadshedding in a segment of Libya’s grid system, aiming to make it imperceptibleto consumers. Ibrahem Alsharif, Hamza Emhemed Hebrisha, and AbdussalamAli Ahmed contribute to the enhancement of quality of life through MI, em-22phasizing the role of Smart Cities in addressing urbanization challenges andpromoting sustainable development in the southern region of Libya. These re-search endeavors highlight the diverse applications of MI in Libya, spanninge-commerce, stock market prediction, energy management, and the creation ofmore livable urban environments. Despite the challenges, these initiatives exem-plify Libya’s commitment to leveraging MI for technological advancement andsocietal well-being.Libya 2020-2023 Concrete ActionsResearch ✓ stock marketTable 4: MI in Libya2.3.1 ResearchSeveral research works have explored MI applications in Libya. [68] AhmedNaji & Ahmed Abu Aeshah investigates MI applications in the adoption of e-commerce in small and medium enterprises (SMEs) in Libya, as presented inthe Journal of University Studies for Inclusive Research. The focus is on MIin Libyan SMEs amid globalization and recent technological revolutions. Theresearch emphasizes the integration of technology into various businesses, par-ticularly small and medium-sized enterprises, providing them with advancedtechnical means for operations and digital transformation. MI proves beneficialfor SMEs in marketing, finance, data capture, employee relationships, and otherbusiness domains. The study aims to examine how MI expands business oper-ations in SMEs, exploring variables influencing its acceptance and the realityof MI and e-commerce in developing countries, with a specific focus on Libya.[69] Najeb Masoud’s research in the British Journal of Economics, Manage-ment & Trade presents techniques and indicators of an Machine neural networkmodel for predicting the movements of the daily Libyan Stock Market (LSM)index. The study covers the period from January 2, 2007, to March 28, 2013,using data from the emerging market of the Libyan Stock Market as a casestudy. Twelve technical indicators serve as inputs for the proposed models, andthe forecasting ability of the ANN model is assessed using metrics like MAE,RMSE, MAPE, and R2. The results indicate that the ANN model accuratelypredicts movement direction with an average prediction rate of 91%, showcasingits effectiveness in forecasting daily stock market prices. The study concludesthat Machine neural networks can serve as a superior alternative technique forpredicting daily stock market prices based on the strong relationship observedbetween parameter combinations and forecast accuracy measures. In [70] AliAA Alarbi, Dani Strickland, Richard Blanchard delve into MI concepts for De-mand Side Shedding Management in Libya in their research presented at the2019 8th International Conference on Renewable Energy Research and Applica-tions (ICRERA). Their study focuses on MI-enabled demand side management,traditionally concerned with adjusting loads to meet generation and address23stability or other constraints. In situations where load surpasses generation,shedding the load becomes more common. The paper specifically examines asegment of the Libyan grid system experiencing daily load shedding and ex-plores the application of MI concepts to intelligently manage this shedding,aiming to make it imperceptible to consumers. In [71] Ibrahem Alsharif, HamzaEmhemed Hebrisha, Abdussalam Ali Ahmed present research on the enhance-ment of the quality of life using MI in their article published in the AfricanJournal of Advanced Pure and Applied Sciences (AJAPAS). Focusing on theimprovement of living standards, the research emphasizes Smart Cities - urbanregions leveraging advanced technology to enhance resident quality of life andoptimize city operations for sustainability. Smart Cities integrate technologiessuch as the Internet of Things, big data analysis, and MI for effective urbanplanning, transportation, energy management, and public services. The au-thors underline the significance of smart cities in addressing the challenges ofurbanization and promoting sustainable development in the southern region ofLibya, aiming to create more livable and resilient environments by enhancinghealthcare services, promoting energy efficiency, reducing traffic congestion, andencouraging citizen participation in civic decision-making processes.2.4 MoroccoIn Morocco, MI has been explored through various research works, highlight-ing its potential to drive progress across sectors like agriculture, healthcare,financial services, and public services. Challenges and opportunities in utilizingdata and MI have been extensively studied, emphasizing the need for decisivepolicy responses to address issues like network limitations, educational readi-ness, and data availability. The impact of MI on human rights is considered,advocating for ethical frameworks guided by principles such as transparency,equity, safety, accountability, and inclusiveness. Practical applications of MIinclude automating water meter data collection for sustainable water use andenhancing decision-making, as well as forecasting regional tourism demand us-ing hybrid models that outperform traditional and MI-based methods. Educa-tional sovereignty and challenges arising from MI are scrutinized, emphasizingthe importance of a national cloud computing structure for safeguarding digitalsovereignty. In industries, startups like Annarabic, SYGMA.AI, Virtual Box,and ATLAN Space showcase Morocco’s innovative strides in speech recognition,insurance claim settlements, VR development, and environmental monitoring.The government actively supports MI development through initiatives like UN-ESCO’s ethics recommendations, the establishment of MoroccoAI, and hostingthe MI Summer School to nurture future MI professionals. Morocco’s com-mitment to MI education and research is further exemplified by the MoroccanInternational Center for MI, fostering expertise in MI and data sciences. Theseendeavors collectively highlight Morocco’s evolving landscape in embracing andleveraging MI for societal advancement and technological innovation. .24Morocco 2020-2023 Concrete ActionsResearch ✓ AI MovementSMB ✓ MoroccoAI, An-narabic, Sigma.AI,AgriEdgeInformalEconomy✓Government✓ Maison de l’intelligenceartificielle Oujda , Mo-roccan InternationalCenter for ArtificialIntelligenceTable 5: MI in Morocco2.4.1 ResearchSeveral research works have explored MI applications in Morocco.The authors of [72] present the challenges and opportunities for developingthe use of data and MI in Morocco. MI has the potential to drive progress, de-velopment, and democratization if governments adeptly handle the challenges.It can significantly enhance productivity growth by extending opportunities incrucial African development sectors such as agriculture, healthcare, financialservices, and public services. MI holds the promise of enabling employees, en-trepreneurs, and enterprises to compete globally and spearhead economic devel-opment by providing access to high-quality digital tools. However, addressingthe accompanying roadblocks requires decisive policy responses. The imple-mentation of MI will necessitate significant adjustments for workers, employers,and businesses, opening new ethical questions that demand thoughtful answers.Specific constraints in Africa, including network limitations, educational institu-tion readiness, and the availability of digital data, further compound the ethicalissues. Aggressive efforts are essential for Africa to overcome these challenges,and success will position the continent to catch up with nations that have al-ready taken strides in MI development. Although the path ahead is intricate,the government’s effectiveness will be measured by its capacity to facilitate col-laboration among all stakeholders, including state and civil society, academics,industry, and national and international entities.The work [73] introduces the topic of MI’s impact on the enjoyment of humanrights and proposes initial considerations for a framework at the national level.This paper primarily comprises a literature review on the subject, accompaniedby concrete recommendations for overseeing MI in Morocco. It advocates forextensive consultations among stakeholders, aiming for co-regulation that cul-minates in the development of an ethical code. This code, guided by a humanrights-based approach, is designed to address key principles, including Trans-parency & Trust, Equity, Safety, Human freedom & autonomy, Accountability25& Justice, Dignity & Integrity, Sustainability, and Solidarity & Inclusiveness.The research presented in [74] centers on the development of a fully MI-based system for automating water meter data collection in Morocco. As thedemand for water resources continues to rise, monitoring becomes crucial forthe rational and sustainable use of this vital resource. Currently, water meterdata collection in Morocco is predominantly performed manually once a monthdue to cost and time constraints. This manual approach often leads to inaccu-rate estimations and calculations, resulting in customer disputes over inflatedinvoices. The paper proposes a comprehensive MI-based system for automatingwater meter data collection, comprising a Recognition System (RS) and a webservices platform. This framework offers a range of services for both customersand water service providers, including consumption monitoring, leak detection,visualization of water consumption, and potable water coverage on a geographicmap. Additionally, it serves as a robust tool for facilitating accurate decision-making through multiple reporting services. The primary component of the RSis a Convolutional Neural Network model trained on a proposed MR-AMR (Mo-roccan Automatic Meter Reading) dataset, achieving an impressive accuracy of98.70% during the model test phase. The system underwent thorough test-ing and validation through experiments. The research outlined in [75] exploresthe forecasting of regional tourism demand in Morocco, comparing traditionaland MI-based methods to ensemble modeling. Tourism stands as a key eco-nomic contributor to Moroccan regions, accounting for 7.1% of the total GDPin 2019. However, this sector remains highly susceptible to external shocks suchas political and social instability, currency fluctuations, natural disasters, andpandemics. To mitigate these challenges, policymakers employ various tech-niques to forecast tourism demand for informed decision-making. The studyspecifically forecasts tourist arrivals to the Marrakech-Safi region using annualdata from 1999 to 2018. Three conventional approaches (ARIMA, AR, and lin-ear regression) are contrasted with three MI-based techniques (SVR, XGBoost,and LSTM). Hybrid models, combining both conventional and MI-based meth-ods through ensemble learning, are then developed. The results reveal thatthese hybrid models outperform both conventional and MI-based techniques,showcasing their ability to overcome individual method limitations.[76] examines Educational Sovereignty and challenges posed by MI in Mo-rocco. The study offers a concise and focused analysis of the primary threatsto educational sovereignty in Morocco in the era of MI. It sheds light on theconcept of educational sovereignty within the Moroccan media and politicaldiscourse surrounding ”ministries of sovereignty.” The article outlines key ini-tiatives and projects in Morocco aimed at overcoming challenges to the educa-tional system and sovereignty in the age of MI, recognizing these systems asindispensable tools in learning and classroom practices. The study primarilydelves into the impacts of using foreign languages, the proliferation of foreignschools and transnational universities, the consequences of Moroccan educatorsand learners extensively employing foreign EdTech, and the influence of plat-forms and programs by major tech companies like GAFAM (Google, Apple,Facebook, Amazon, and Microsoft). It also addresses the threats arising from26the absence of a national cloud computing structure, essential for safeguard-ing digital sovereignty and protecting the personal data of learners and theeducational system. The article contends that despite Morocco’s aspirationsto establish a digital ecosystem capable of preserving educational sovereignty,numerous subjective and objective obstacles still impede these plans.The study presented in [77] employed factor analysis, a confirmatory methoddesigned to identify latent factors from observable variables. This method in-volves assigning a set of observable characteristics to each latent variable, and bysetting parameters (loadings) to 0 in confirmatory factor analysis, further analy-sis becomes possible, allowing correlations between latent factors and, if needed,additional correlations between residual errors. This process offers a comprehen-sive description of hidden variables. The central question posed by the authorswas: To what extent does digitalization contribute to reducing inequalities intechnological acceptance in Morocco? Confirmatory factor analysis, a special-ized form of structural equation modeling, was used. In this approach, a modelis predefined, specifying the number of factors, potential relationships betweenthese factors, connections between these factors and the observed variables,error terms associated with each observed variable, and possible correlationsbetween them. The sample focused on stakeholders of the Digital DevelopmentAgency, with a selected group of 60 stakeholders, comprising 20 companies,20 associations, and 20 cooperatives. Utilizing the Unified Theory of Accep-tance and Use of Technology (UTAUT) model, the study provides estimatesindicating that ease of use, quality of service, anticipated performance, and theinfluence of the professional body are all concepts contributing to psychologicaland motivational acceptance for the use of MI.The research presented in [78] explores the opportunities and challengesassociated with integrating MI into International Financial Reporting Standards(IFRS) within accounting systems. Focused on the context of Morocco, thepaper provides insights into various aspects of finance and accounting in thecountry. It addresses challenges related to adoption, education, and technicalexpertise in Morocco, alongside examining the ethical and legal implications ofMI-based accounting systems. Their article further presents into the importanceof maintaining compliance with IFRS requirements and ensuring the integrityand transparency of financial data. The authors conclude that a comprehensiveunderstanding of IFRS and the adoption of MI in accounting systems are crucialfor effectively navigating the challenges and opportunities presented by thesedevelopments.2.4.2 Small BusinessesAnnarabic is an MI startup in Morocco that aims to empower every Arab voice.Their focus areas include MI Software API, Customer Satisfaction, AutomaticSpeech Recognition (ASR), Customer Feedback, and Natural Language Process-ing. The startup is actively involved in developing speech recognition systemsfor Arabic dialects, offering services such as audio transcription for call centers,audio intelligence for call centers, retail, and social media companies (covering27sentiment analysis, keyword flagging, etc.), as well as Chatbot/Voicebot solu-tions to facilitate customer navigation and order processing, and subtitling andcaptioning for videos. SYGMA.AI revolutionizes the insurance industry by sig-nificantly reducing the time required to settle car accident claims. With anaverage settlement time of less than 5 minutes, the fully automated end-to-endSAAS solution utilizes MI visual inspection to detect car damages, estimaterepair costs, and manage claims based solely on mobile photos. SYGMA.AI’scapabilities extend to eliminating fraud, thereby saving valuable time and moneyfor insurance companies and their clients. Virtual Box, located in the heart ofCasablanca, is an independent multidisciplinary studio dedicated to creatingand developing Virtual Reality applications, interactive maps, and immersiveexperiences utilizing 3D, 360 Video, and Photography/VR. AgriEdge offers adata-driven decision support platform for crop production management. Lever-aging data from satellites, drones, field sensors, weather, and market prices,it provides farmers with decision proposals related to strategic processes likeirrigation, fertilization, plant disease management, and yield prediction. Theplatform delivers these proposals through user-friendly mobile and web applica-tions, specifying the right place, time, and quantity. Moroccan startup ATLANSpace has secured 1.1 million USD in Series A funding to further develop its MIsystem guiding unmanned aircraft on data collection and tracking missions overlarge geographical areas. The investment, led by Maroc Numeric Fund II, buildson the startup’s initial seed funding in 2019. ATLAN Space’s unique technol-ogy, recognized globally, supports governments and institutions in combatingenvironmental crimes like illegal fishing or deforestation. Founded in 2016 byBadr Idrissi, ATLAN Space has received acclaim, including recognition fromNvidia as one of its top ten MI startups in 2018. DataPathology is a Moroccanmedtech startup that provides remote pathology consultation.2.4.3 GovernmentIn terms of development and adoption of MI, Morocco has made some progress,mainly through the establishment of academic institutions, hosting internationalconferences, implementing effective training and educational programs, and set-ting up large-scale data centers.In December 2018, UNESCO and the Polytechnic University Mohamed VIorganized their very first important International forum on Artificial Intelligencein Africa, in Benguérir, Morocco. The Forum’s objective was to enrich theglobal reflection on artificial intelligence by drawing up a complete inventoryof the situation and towards an assessment of an African scale, by taking intoaccount challenges, opportunities and issues specific to local contexts. About150 participants representing the Member States as well as high-level public andprivate sector partners participated to this event.Morocco stands out as one of the earliest countries to implement UNESCO’srecommendations on MI ethics, adopted during the 41st session of UNESCO’sGeneral Conference in November 2021 in Paris. Established in 2019, MoroccoAIis a prominent Non-Governmental Organization , spearheaded by esteemed MI28scientists and researchers in Morocco and abroad. MoroccoAI core focus is onpromoting MI education and fostering excellence in research and innovation,in Morocco and across the broader African landscape and beyond. The MISummer School offers an immersive and transformative experience, empoweringparticipants to excel in these technologies. The organizers, Al Akhawayn Uni-versity and MoroccoAI, come together in a strategic partnership through thisdistinguished summer school to foster a highly proficient workforce and culti-vate the next generation of MI professionals and leaders. The first MI SummerSchool of Morocco was held in Ifrane Morocco from July 17th-21st, 2023. MImovement, the Moroccan International Center for Artificial Intelligence is a cen-ter of excellence in MI that aims to foster the emergence of Moroccan expertisein MI and Data Sciences.On the Knowledge Campus, Oujda, Maison de l’intelligence artificielleMIA-UMPO-Maroc is designed to better understand MI and its transformations.MI opens up new avenues for the dissemination of expert and ethical knowl-edge to the general public while providing opportunities to develop innovativeand collaborative projects by bringing together various stakeholders in the MIecosystem. Data, artificial intelligence, and the Internet of Things are presentin our lives, reshaping our relationship with knowledge and the economy, po-tentially reshaping our future and transforming our societies. That’s why theMohammed Premier University, like the MIA in Sophia Antipolis, wanted tocreate a space where reflections and experiments related to AI focus on unit-ing all stakeholders around a future shaped by MI. MIA-UMPO-Maroc aims toinvigorate institutional, academic, and industrial collaboration around new tech-nologies and their challenges. The space will be regularly animated by events oninnovation and AI, allowing the public to learn, exchange, and develop knowl-edge. MIA-UMPO-Maroc aims to create a real dynamic of AI acculturationthrough public experiences and the promotion of applied research with a signif-icant societal and economic impact. MIA-UMPO-Maroc consists of an EthicalWatch Observatory, Exhibition Hall, Coworking Space, and ExperimentationLaboratory.2.5 SudanIn Sudan, the deployment of MI from 2000 to 2023 has been marked by im-pactful research initiatives and innovative applications. MI played a crucialrole in assessing groundwater quality in northern Khartoum State, employingmultilayer perceptron neural networks and support vector regression to modelsuitability for drinking. The models demonstrated efficiency, showcasing theirpotential in groundwater quality evaluation. Additionally, MI addressed healthchallenges in Gedaref State, forecasting malaria and pneumonia cases usingARIMA and Prophet models. The industry sector witnessed the emergence ofATLAN Space’s MI-enabled drones for environmental conservation, contribut-ing to efforts like combatting desertification. The SudanAI model, a collabora-tive venture between local researchers and global MI organizations, exemplifiesSudan’s commitment to creating a cutting-edge MI solution tailored to the nu-29ances of the Arabic language spoken in Sudan. These endeavors collectivelyhighlight Sudan’s stride towards leveraging MI for research, environmental con-servation, and linguistic adaptation, showcasing a growing engagement withtechnological advancements to address societal challenges.Sudan 2020-2023 Concrete ActionsResearch ✓ SudanAISMB ✓ ATLANGovernment✓Table 6: MI in Sudan2.5.1 ResearchThe work in [79] investigated groundwater quality in northern Khartoum State,Sudan, utilizing MI algorithms. The authors employed multilayer perceptronneural network and support vector regression to assess groundwater suitabilityfor drinking. The groundwater quality was evaluated through the groundwaterquality index (GWQI), a statistical model using sub-indices and accumulationfunctions to reduce data dimensionality. In the first stage, GWQI was calculatedbased on 11 physiochemical parameters from 20 groundwater wells, indicatingthat most parameters exceeded World Health Organization standards, exceptEC and NO3-. The GWQI ranged from 21 to 396, classifying samples intoexcellent (75%), good (20%), and unsuitable (5%) water categories.To overcome computational challenges, the study applied MI techniques,specifically MLP neural network and SVR models. The models were trainedand validated on a dataset divided into 80% for training and 20% for vali-dation. Comparison of predicted and actual (calculated GWQI) models us-ing MSE, RMSE, MAE, and R2 criteria demonstrated the robustness and effi-ciency of MLP and SVR models. Overall, groundwater quality in north Khar-toum is deemed suitable for human consumption, except for BH 18, which ex-hibits highly mineralized water. The developed approach proves advantageousfor groundwater quality evaluation and is recommended for incorporation intogroundwater quality modeling.The work in [80] centers on Endemic Diseases: A Case Study of GedarefState in Sudan, employing MI technologies. Smart Health, a crucial compo-nent, enhances healthcare through services like disease forecasting and earlydiagnosis. Although numerous machine learning algorithms support S-Healthservices, the optimal choice for disease forecasting remains uncertain. GedarefState faces persistent challenges with malaria and pneumonia due to heavy rain-fall. Predicting future trends in these diseases is vital for effective preventionand control. This study utilizes a time series dataset from the state’s ministry ofhealth to estimate malaria and pneumonia cases in Gedaref State, Sudan, fivemonths later. Two forecasting methodologies, ARIMA and Prophet models,are applied, comparing their performance in predicting diseases. Data collected30from January 2017 to December 2021 reveals that the ARIMA technique outper-forms FB-Prophet in forecasting both malaria and pneumonia cases in GedarefState.2.5.2 Small BusinessesMI-enabled Drones for good. ATLAN Space is one of many startups innovat-ing with drone technology on the continent. Across Southern Africa, dronesare used to protect elephants and rhinos from poaching. In Sudan, a startupwants to drop Acacia tree seeds from the sky to tackle desertification, and inSouth Africa, drones are used in agriculture to monitor crop health and detectdisease. The Sudanese MI model, also known as the SudanAI, is an MI modelthat is specifically designed to understand and process the Arabic language spo-ken in Sudan, as well as capture the nuances of Sudanese dialects and culturalreferences. The development of the Sudanese MI model is a collaborative ef-fort between local Sudanese researchers, engineers, and data scientists, who areworking in partnership with global MI organizations to create a cutting-edgeMI solution that addresses the unique needs and challenges of Sudan.2.6 TunisiaIn Tunisia, the integration of MI has been instrumental in diverse research en-deavors and strategic government initiatives from 2015 to 2023. MI empowersthe Tunisian textile industry by providing a competitive edge through informeddecision-making, as outlined in [83]. Additionally, the application of low-powerblockchain and MI in water management exemplifies Tunisia’s Industry 4.0endeavors, enhancing services and reinforcing trust among stakeholders [84].Research efforts address environmental challenges, such as forecasting urbansolid waste using sequential MI models, emphasizing the effectiveness of LSTMand bidirectional LSTM in predicting optimal waste bin numbers [85]. On thelinguistic front, Tunisia advances MI through the creation of an end-to-endspeech recognition system for the Tunisian dialect, showcasing innovative appli-cations in under-resourced languages [86]. The technology industry is booming,marked by Instadeep’s acquisition, highlighting Tunisia’s emergence as an en-trepreneurial and innovative hub in deep tech, attracting global attention [83].In the governmental sphere, Tunisia’s commitment to reform includes incor-porating MI in Public Finance Management Information Systems, enhancingtransparency and accountability to rebuild public trust [81], [82]. These collec-tive efforts underscore Tunisia’s strategic use of MI across sectors for innovation,environmental sustainability, and governance.2.6.1 Research[83] aims to empower the Tunisian textile industry with MI. Textile holdsparamount importance in Tunisia, being the leading industry in terms of em-ployment and the number of companies. It is crucial for the country to stay31Tunisia 2020-2023 Concrete ActionsResearch ✓ TunSpeech, textile,Afro-Mediterraneanmeeting in the field ofArtificial IntelligenceIndustry InstadeepSMB ✓ KatYos, WARM,ScorSell, LWATN,AD’VANTAGE, Busi-ness & AIInformalEconomy✓ Deepera.AIGovernment✓ PFMIS, National MIStrategyTable 7: MI in Tunisiaabreast of the latest technologies to maintain global competitiveness. MI is iden-tified as a technology providing a competitive advantage, aiding sector leadersin making informed business and strategic decisions. The article introduces MIin the context of the Tunisian textile industry, outlines use cases, and providesrecommendations to industry stakeholders. The article maintains a deliberatelyhigh level to be accessible to a diverse audience.[84] focuses on low-power blockchain to study water management in Tunisiawithin the context of Industry 4.0. The industrial sector is evolving towards In-dustrial IoT (IIoT) and Industry 4.0, where blockchain technology can addresslimitations related to security and data reliability in IoT. The article presentsa new platform integrating MI and smart contracts to monitor and track waterconsumption in Tunisia. The proposed solution enhances water managementservices, offering benefits such as consumption monitoring, traceability, secu-rity, water leak detection, and visualization of water consumption and drinkingwater coverage. This approach aims to strengthen trust and security amongvarious stakeholders. [85] utilizes sequential MI models to forecast urban solidwaste in the city of Sousse, Tunisia. Urban solid waste poses a significantenvironmental challenge, with waste generation linked to economic growth, in-dustrialization, urbanization, and population growth. The article focuses onpredicting solid waste generation based on monthly recorded waste amounts todetermine the optimal number of waste bins. Various MI regression and clas-sification approaches are evaluated, highlighting the effectiveness of sequentialmodels—specifically, long short-term memory and bidirectional LSTM - in pre-dicting the number of waste bins compared to other methods. [86] focuses onTunisian dialectal end-to-end speech recognition using deepspeech. Automat-ically recognizing spontaneous human speech and transcribing it into text isan increasingly important task. However, freely available models are scarce,32especially for under-resourced languages and dialects, as they require extensivedata to achieve high performance. The paper outlines an approach to construct-ing an end-to-end Tunisian dialect speech system based on deep learning. Toachieve this, a Tunisian dialect paired text-speech dataset named ”TunSpeech”was created. Existing Modern Standard Arabic (MSA) speech data was com-bined with dialectal Tunisian data, reducing the Out-Of-Vocabulary rate andimproving perplexity. Conversely, the introduction of synthetic dialectal datathrough text-to-speech increased the Word Error Rate.2.6.2 IndustryTunisia, a North African country, is making significant strides in the technologyindustry. With a rapidly evolving entrepreneurship and innovation ecosystem,Tunisia is gradually establishing itself as a leading destination for entrepreneur-ship and innovation, especially in deep tech. The acquisition of Instadeep, a deeptech startup founded in 2014 by Tunisians Karim Beguir and Zohra Slim, hassent shockwaves through the Tunisian entrepreneurship and innovation ecosys-tem. BioNTech acquired Instadeep for 362 million euros upfront, with an addi-tional 200 million euros contingent on future performance. Such acquisitions arepivotal milestones for developing ecosystems. Instadeep’s success underscoresthe potential of Tunisian startups to innovate, compete globally, and attractattention from major players in the tech industry.IEEE AMCAI 2023 was organized by ATIA (Association Tunisienne pourl’Intelligence Artificielle) with the Financial Co-Sponsorship of IEEE AfricaCouncil, and the Technical Co-Sponsorship of IEEE Tunisia Section, IEEE SPSTunisia Chapter, IEEE CIS Tunisia Chapter, IEEE SMC Tunisia Chapter inDecember 2023, at Hammamet, Tunisia. IEEE AMCAI’2023 is the premierAfro-Mediterranean meeting in the field of Artificial Intelligence. Its purpose isto bring together researchers, engineers, and practitioners from both Mediter-ranean and African countries to discuss and present their research results aroundArtificial Intelligence and its applications. The conference emphasizes appliedand theoretical researches in MI to solve real problems in all fields, includingengineering, science, education, agriculture, industry, automation and robotics,transportation, business and finance, design, medicine and biomedicine, bioin-formatics, human-computer interactions, cyberspace, security, Image and VideoRecognition, agriculture, etc.2.6.3 Small BusinessesFounded in 2019, ScorSell, Mobile classifieds that Inspires everyone in the worldto start selling. It provides a highly scalable e-commerce platform based on newtechnologies such as machine intelligence , machine learning , image recognition, data science . With a very friendly user experience and attractive design,ScorSell aims to create a secure payment system based on smart contract tech-nology with low fees.33LWATN leads the legal tech frontier in Tunisia with its state-of-the-art MI-powered chatbot. This groundbreaking tool not only facilitates seamless con-versations and delivers precise legal guidance with a clear voiceover in variouslanguages, but also prioritizes accessibility.AD’VANTAGE is a startup that comes up with an innovative idea whichwould rethink classical loyalty programs used by companies to ensure their cos-tumers faithfulness.WARM is an online marketplace that helps connect buyers and sellers tofind the right piece of furniture and decor for their homes and offices. WARM isconceived as a smart platform that answers our current furniture consumptionproblems that might be a frustrating experience. Lack of online presence forfirsthand brands or dealing with strangers and organizing heavy pickups forsecondhand products.KatYos connects opticians and eyewear buyers by offering a 100% digitalexperience, from online fitting to corrective lens processing. It is an optometrymarketplace that allows opticians to digitize their business easily and withoutinvestment by offering an immersive, unique, safe and secure shopping expe-rience to their customers with the help of an enhanced digital mirror, wherecustomers can evaluate a potentially unlimited number of frames anywhere andanytime.In Tunisia, Business & AI innovates with MI-powered decision-making solu-tions to help companies improve operational processes and make informed deci-sions. Save Your Wardrobe is a FashionTech startup using AI to make Fashionmore digital and sustainable. Wattnow help compagnies gain actionable insightson their overall energy usage.2.6.4 Informal EconomyDeepera.AI in Tunisia develops smart solutions for investors and stock traders,incorporating MI into financial tasks and offering products like stock exchangemanagement tools.2.6.5 GovernmentIn 2015, the Tunisian government initiated a series of public sector reformsto enhance government operations and address citizen needs. Recognizing thenecessity for continued reforms, it is advocating for the implementation of anaccountable Public Finance Management Information System (PFMIS) and pro-poses introducing MI into the current financial system, identified as a high-riskarea for corruption. In 2016, the government outlined its vision for MI, alongwith other priorities, in a strategy document detailing a five-year developmentplan for Tunisia (2016-2020). This strategy was later complemented by the gov-ernment’s economic and social roadmap for 2018-2020. The economic and socialroadmap for 2018-2020 aims to expedite the reforms initiated under the earlierfive-year development plan. The overarching goal of the development plan isto ensure human rights, social well-being, and economic growth in Tunisia. To34formulate its National MI Strategy, the Secretary of State for Research estab-lished a Task Force in 2018 to oversee the project and a Steering Committeeto develop a methodology and action plan for the strategy. To kick off thisnational initiative, the UNESCO Chair on Science, Technology, and Innova-tion Policy, in collaboration with the National Agency for Scientific ResearchPromotion-ANPR, will host a workshop titled “National MI Strategy: Unlock-ing Tunisia’s capabilities potential.” The primary goal of the gathering was toshare and discuss the proposed framework, methodology, and action plan putforth by the Steering Committee for designing the strategy. The event tookplace at ENIT, specifically in the Mokhtar Latiri Amphitheater. Concrete MIapplications in the Public Finance Management Information System (PFMIS)has conducted in Tunisia. The PFMIS comprises core subsystems that furnishthe government with essential information for planning, executing, and moni-toring public finances. Its scope and functionality encompass fraud detection,budget efficiency, and financial analytics. By employing a combination of ma-chine learning, big data, and natural language processing techniques, MI aidsauditors and finance officials at the Ministry of Finance in managing the vastamounts of data essential for meeting transparency and accountability require-ments in fulfilling fiduciary responsibilities to Tunisian taxpayers. This, in turn,contributes to rebuilding public trust in the government [81], [82].3 Southern AfricaIn Southern Africa, MI activities are diverse.Botswana explores water billing, diabetic retinopathy screening, solar radi-ation prediction, gravel loss condition prediction, HIV/AIDS treatment, datamining, ART program success, and clinical decision support. Eswatini engagesin financial inclusion, diarrhoea outbreak prediction, renewable energy, COVID-19 case prediction, invasive plant study, MI in education, maize crop yield pre-diction, and economic development through technology. Lesotho integrates MIin land cover mapping, carbon sequestration, soil organic carbon prediction,weather nowcasting, health insurance enrollment, education, electricity demandforecasting, legal frameworks, wage impact, industry mergers, and healthcarepolicy.Namibia’s MI landscape is equally dynamic, exploring MI in education, pre-dicting Gender-Based Violence through machine learning, and examining cy-bersecurity practices in rural areas. Other endeavors include using MI to dis-criminate individual animals with tags in camera trap images and developinga smart irrigation system for efficient farming. The UNESCO Southern Africasub-Regional Forum on Artificial Intelligence in Windhoek further emphasizesthe commitment to sustainable and ethical MI use in the region.In South Africa, Accenture collaborates with the Gordon Institute of Busi-ness Science to guide businesses through digital transformation, emphasizing theintegration of MI in Public Sector Human Resource Management and MI-basedmedical diagnosis for improved healthcare.35The pervasive use of MI across these regions reflects its transformative im-pact in various domains.3.1 BotswanaBotswana 2020-2023 Concrete ActionsResearch ✓ waterSMB ✓InformalEconomy✓Government✓Table 8: MI in Botswana3.1.1 Research[132] Gaboalapswe’s research explores an MI and data analytics model to ad-dress domestic water billing crises in Botswana. The study integrates MI andBig Data Analytics for efficient household water meter reading and proposesan enterprise integration system for water utility corporations. [133] Kgame,Wu, and Geng investigate eye care physicians’ knowledge and perceptions ofusing MI in screening for diabetic retinopathy in Botswana. The study revealslimited awareness of MI technology among clinicians but positive attitudes to-ward its potential in diabetic retinopathy screening. [134] Sampath Kumar,Prasad, Narasimhan, and Ravi discuss the application of machine neural net-works for predicting solar radiation in Botswana. The study utilizes data frommultiple locations and encourages the development of a significant model forestimating solar potential in the country. [135] Oladele, Vokolkova, and Eg-wurube apply a Feed Forward Neural Network to predict gravel loss conditionfor low volume road networks in Botswana. The FFNN model proves accuratein predicting gravel loss, providing valuable information for future maintenanceinterventions. [136] Dr. Chandrasekaran focuses on clustering as an MI tech-nique in drug resistance of HIV/AIDS patients in Botswana. The study aims toaddress challenges in the Messiah ARV program and improve the effectivenessof HIV/AIDS treatment. [137] Anderson, Masizana-Katongo, and Mpoelengexplore the potential of data mining in Botswana. The paper discusses howorganizations in Botswana can benefit from data mining technologies to trans-form raw data into valuable information for strategic decision-making. [138]Smartson and Thabani Nyoni analyze Botswana’s ART program success usingmachine neural networks. The study employs an ANN model to forecast upwardtrends in annual ART coverage from 2019 to 2023, emphasizing the importanceof improving ART access. [139] Ndlovu et al. evaluate the feasibility and ac-ceptance of a mobile clinical decision support system (VisualDx) in Botswana.36Healthcare workers express interest in using mHealth technology, highlightingthe potential benefits of VisualDx in resource-constrained environments.3.2 EswatiniEswatini 2020-2023 Concrete ActionsResearch ✓ eSwatiniSMB ✓InformalEconomy✓Government✓Table 9: MI in Eswatini3.2.1 Research[140]) explores the application of machine learning on financial inclusion datain Eswatini, focusing on small-scale businesses. Contributes to understandingthe extent of financial inclusion and the challenges faced by entrepreneurs. [141]develops a machine learning-based surveillance system for predicting diarrhoeaoutbreaks in Eswatini. The study employs supervised learning models and high-lights the effectiveness of Support Vector Machine (SVM) in data classification.[142]) focuses on classifying financial inclusion datasets in Eswatini using SVMand logistic regression. Recommends attention to enhancing financial inclusion,particularly in Hhohho, Shiselweni, and Lubombo regions, with mobile moneyas a key indicator. [143] presents research on renewable energy sources in Eswa-tini, emphasizing the need for alternative energy to supplement existing hydropower. Aims to contribute to the country’s self-sufficiency in power genera-tion. [144] utilizes Machine neural networks to predict daily COVID-19 cases inEswatini, providing insights into disease forecasting. The study emphasizes theimportance of adhering to WHO guidelines for disease prevention and control.[145] applies machine-learned Bayesian networks to study the invasion dynamicsof the highly invasive plant, Chromolaena odorata, in Eswatini. Contributes tounderstanding factors influencing the species’ geographic distribution patterns.[146] proposes a theoretical framework for integrating MI into Library and In-formation Science curricula at the University of Eswatini. Contributes to theongoing discourse on incorporating MI in education. [147] addresses the criticalissue of predicting maize crop yields in Eswatini using machine learning, pro-viding insights into how this technology can contribute to alleviating povertyand ensuring food security. [148] highlights the work of DataNet in eSwatini,showcasing how the organization contributed to the country’s economic devel-opment through innovative strategies and technology implementations, such asthe National Fire and Emergency Response System.373.3 LesothoLesotho 2020-2023 Concrete ActionsResearch ✓ TradeSMB ✓ BasaliTech, Plano-OneTree, eFarmers,LibraChat, Mphalane,Keti, Nalane, QoloqoInformalEconomy✓Government✓Table 10: MI in Lesotho3.3.1 ResearchThe work in [170] focuses on the integration of machine learning and open-accessgeospatial data for land cover mapping. The authors developed a frameworkusing 10-m resolution satellite images and machine learning techniques for landcover mapping in Lesotho. This work highlights the use of MI for precise map-ping in developing countries.In [171], the study addresses carbon sequestration potential in Lesotho’scroplands. Through the application of a quantile random forest model, theresearch emphasizes the role of sustainable soil management practices in re-plenishing carbon stocks. The AI-driven model aids in determining soil organiccarbon content, showcasing the integration of MI for environmental assessment.The paper [172] presents an innovative methodology for operational landcover mapping in Lesotho using Earth Observation data. The study utilizesmachine learning techniques, specifically a Random Forest Classifier, for gener-ating standardized annual land cover maps. This demonstrates the effective useof MI in overcoming challenges related to limited in-situ data, a common issuein developing countries.In [173], the focus is on predicting soil organic carbon content using hyper-spectral remote sensing in a mountainous region of Lesotho. The study employsmachine learning algorithms, specifically random forest regression, to accuratelyestimate soil organic carbon. This work showcases the utility of MI in predictingsoil properties for agricultural and environmental purposes.The work by Senekane et al. in [174] explores weather nowcasting using deeplearning techniques. The study employs multilayer perceptron, Elman recurrentneural networks, and Jordan recurrent neural networks for short-term weatherforecasting. The high accuracies achieved highlight the effectiveness of MI inpredicting weather variations.In [175], the authors utilize machine learning techniques to predict healthinsurance enrollment and analyze the take-up of health services in sub-Saharan38Africa. The study demonstrates that applying different machine learning ap-proaches improves the identification of excluded population groups, emphasizingthe role of MI in enhancing health policy research.The work in [176] investigates students’ intention to learn MI in Lesotho’ssecondary schools. The study identifies factors influencing students’ intentions,emphasizing the impact of attitudes, confidence, self-efficacy, and subjectivenorms. This work underscores the importance of creating a supportive environ-ment for MI education and highlights AI’s role in shaping educational intentions.In [177], the study focuses on short-term electricity demand forecasting forLesotho. Utilizing ABB Nostradamus software, the research achieves accurateday-ahead, week-ahead, and hour-ahead electricity demand forecasting results.The study recommends bilateral contracts and the use of AI-based forecastingfor optimal power procurement, showcasing the significance of MI in energyplanning.The work by Ramokanate in [178] critically examines the approach under theLesotho Electronic Transactions and Electronic Commerce Bill 2013, specificallyregarding the use of electronic agents in trade and commerce. The analysisrecommends considering the use of the law of agency for instances involvingelectronic agents. This emphasizes the need for legal frameworks to adapt toadvancements in MI.In [179], the study explores the impact of MI on average wages in South-ern Africa. Using a panel VAR approach, the research reveals a significantnegative relationship between MI and average wages. The study suggests pol-icy directions focusing on wage stabilization, income redistribution, and skilldevelopment to address the effects of MI on wages.The article in [180] investigates knowledge retention in the cross-bordermergers of Lesotho’s telecommunications industry. The study identifies chal-lenges and successes in knowledge retention during mergers, emphasizing theneed for formal policies. The research underscores the role of MI in optimizingknowledge management processes during industry transformations.The research by Katende et al. in [181] evaluates the impact of a multi-disease integrated screening and diagnostic model for COVID-19, TB, and HIVin Lesotho. The study utilizes AI-driven screening and diagnostic services, main-taining integrated testing despite disruptions. The findings highlight the posi-tive effects of MI in ensuring timely diagnoses and linkage to healthcare services.In [182], the study predicts health insurance enrollment and analyzes thetake-up of health services in 38 sub-Saharan African countries. Using machinelearning techniques, the research identifies excluded population groups, improv-ing targeting for health policy. This work emphasizes the role of MI in enhancingthe accuracy of predictions and informing healthcare decisions.Across these works, the common theme is the pervasive use of MI in ad-dressing diverse challenges in Lesotho. From educational intentions and energyforecasting to legal frameworks, labor markets, industry mergers, and healthcarepolicy, these studies collectively demonstrate the versatile and transformativeimpact of MI across various domains.393.3.2 Small BusinessesBasaliTech, a Lesotho non-profit that aims to address gender parity in STEMthrough training such as computing, programming( web and mobile), and elec-tronics and robotics for young girls and children. Plant One Tree, a youth-empowered environmental organization that seeks to plant all trees. eFarmersLesotho, Agro mobile App: e-extension and virtual marketing. Mphalane pro-vides automated chat assistants for your business. Mphalane is a company witha product called LibraChat. LibraChat is a communication channel for everyentity. The product is a 24-hour chatbot focused on increasing communication,enquiries, shopping, ordering and many other forms of services that may requirecommunication. Founded in 2018, KETI aims to use MI to improve healthcareoutcomes in Lesotho and beyond. The startup has developed a platform thatuses machine learning algorithms to analyze medical data and identify patternsthat can help doctors make more accurate diagnoses and treatment decisions.Nalane is a social venture that engages in viable social enterprising initiatives toraise resources to sustain its mission of ensuring that public education works forall in Lesotho and South Africa. Through inclusive and comprehensive after-school programs at primary school level, they especially target Orphans andVulnerable Children to increase education attainment rates. Qoloqo focuses onimproving access to financial services for underserved communities. Qoloqo’splatform uses MI to analyze financial data and provide personalized recommen-dations to users, such as which savings accounts or loans would be the bestfit for their needs. The startup has already partnered with several banks andmicrofinance institutions in Lesotho, and is expanding into other countries inthe region3.4 NamibiaNamibia 2020-2023 Concrete ActionsResearch ✓ Gender-Based ViolenceSMB ✓ Tincup, Loop, Hyper-link, AfriFeelGovernment✓ Windhoek StatementTable 11: MI in Namibia3.4.1 Research[233] reports some experiences of teaching MI in a Namibian university in col-laboration with a Finnish university and a few companies. Within the Comput-ing Education Community, only a minority of research reports have experienceteaching MI, and very little research has been conducted on teaching and learn-ing MI in Africa. Given the high importance and impact of MI, this is alarming.40Learning and teaching MI in an African higher education setting provides uniquechallenges compared to the standardized approach in the Global North. Ourundergraduate course in MI was carried out in a novel way that emphasizedthe creative application of MI to meet the requirements of the Fourth IndustrialRevolution (4IR). We chose an approach that helps Computer science graduatesto explore and get inspired by the opportunities of MI at the ground.Gender-Based Violence (GBV) is a pervasive issue that poses a significantthreat to individuals and communities worldwide. Many countries, faces thepressing challenge of combating GBV as an impediment towards ensuring thesafety and well-being of its citizens. Despite the coordinated efforts taken byseveral governments and its people, there remains a cause for concern regard-ing curbing GBV globally. Women, children, and other vulnerable members ofsociety continue to be exposed to this significant threat. Current measures areprimarily reactive, with focus on punishing perpetrators rather than foresee-ing and preventing future occurrences. [234] provides a machine leaning-basedanalytical process for predicting the occurrence of GBV to aid in early preven-tion strategies. A preliminary case study of the analytical process was recom-mended using Namibia as one of the South African countries. External datasetof GBV were extracted from kaggle platform containing some Namibia recordsfor preliminary study and model testing. It was recommended in this studythat Namibia should employ various machine learning algorithms as a test case.These algorithms should be compared and evaluated for their predictive accu-racy in forecasting GBV events in Namibia. Additionally, the case study shouldconduct exploratory data analysis to identify key trends and drivers of GBVin Namibia, providing valuable insights for policymakers and intervention pro-grams. By leveraging advanced ML techniques, the case study would contributeto evidence-based decision-making, policymaking, planning, and resource allo-cation aimed at reducing GBV incidents in Namibia. If the analytical processis duly followed and applied in Namibia, the research outcome would have thepotentials to inform policymakers, law enforcement agencies, and social orga-nizations about the underlying causes and risk factors associated with GBV,enabling them to implement effective intervention strategiesGlobally, Information Communication Technology device usage has seen asteep rise over the last few years. This also holds in developing countries,which have embarked on connecting the unconnected or previously disadvan-taged parts of their populations. This connectivity enables people to interactwith cyberspace, which brings opportunities and challenges. Opportunities suchas the ability to conduct business online, attend online education, and performonline banking activities. Challenges experienced are the cost of Internet accessand more worrying cyber-risks and potential for exploitation. There remainpockets of communities that experience sporadic connectivity to cyberspace,these communities tend to be more susceptible to cyber-attacks due to issuesof lack/limited awareness of cyber secure practices, an existent culture thatmight be exploited by cybercriminals, and overall, a lackluster approach to theircyber-hygiene. The work in [235] presents a qualitative study conducted in ruralNorthern Namibia. Our findings indicate that both secure and insecure cyberse-41curity practices exist. However, through the Ubuntu and Uushiindaism Afrocen-tric lenses, practices such as sharing mobile devices without passwords amongthe community mirror community unity. Practices such as this in mainstreamresearch can be considered insecure. We also propose interrogating “common”secure cybersecurity practices in their universality of applicabilityThe use of technology in ecology and conservation offers unprecedented op-portunities to survey and monitor wildlife remotely, for example by using cameratraps. However, such solutions typically cause challenges stemming from the bigdatasets gathered, such as millions of camera trap images. MI is a proven, pow-erful tool to automate camera trap image analyses, but this is so far largely beenrestricted to species identification from images. The work in [236] develops andtests an MI algorithm that allows discrimination of individual animals carryinga tag (in this case a patagial yellow tag on vultures) from a large array of cameratrap images. Such a tool could assist scientists and practitioners using similarpatagial tags on vultures, condors and other large birds worldwide. We showthat the overall performance of such an algorithm is relatively good, with 88.9%of all testing images (i.e. those not used for training or validation) correctlyclassified using a cut-off discrimination of 0.4. Specifically, performance washigh for correctly classifying images with a tag (95.2% of all positive imagescorrectly classified), but less so for images without a tag (87.0% of all negativeimages). The correct classification of images with a tag was, however, signif-icantly higher when the tag code was at least partly readable compared withthe other cases. Overall, this study underscores the potential of MI for assistingscientists and practitioners in analysing big datasets from camera traps.Farmers in Namibia currently operate their irrigation systems manually, andthis seems to increase labor and regular attention, especially for large farms.With technological advancements, the use of automated irrigation could allowfarmers to manage irrigation based on a certain crops’ water requirements. Thework in [237] looks at the design and development of a smart irrigation systemusing IoT. The conceptual design of the system contains monitoring stationsplaced across the field, equipped with soil moisture sensors and water pumps tomaintain the adequate moisture level in the soil for the particular crop beingfarmed. The design is implemented using an Arduino microcontroller connectedto a soil moisture sensor, a relay to control the water pump, as well as a GSMmodule to send data to a remote server. The remote server is used to representdata on the level of moisture in the soil to the farmers, based on the readingsfrom the monitoring station3.4.2 Small BusinessesAfriFeel Digital Innovations: In today’s fast-paced world, finding the best traveldeals is crucial for both leisure and business travelers. Afrifeel travel app makesit effortless to discover the most up-to-date and exclusive travel bargains.Hyperlink InfoSystem is well known to craft the most innovative & eyecatchy mobile apps & websites. It offers a wide range of customized services inmobile apps, website development, AR-VR Development, Game Development,42Blockchain Development and much more.Loop Technologies deliver exceptional experiences that drive your businessforward. Floor standing machines, tablets, signage, selfservice data collectionand more.TinCup Digital: TinCup was established in order to facilitate the deliveryof integrated digital and marketing services to clients in Namibia, clients whountil now have had to be content with existing above-the-line strategies dictat-ing the direction of integrated campaigns. We aim to rewrite the advertisingand marketing handbook in Namibia. With decades of combined experience inthe Namibian digital, marketing and advertising landscape, we offer clients aholistic view of their brands in real life, from the first point of contact throughto campaign conceptualisation, development, integration and alignment withcurrent strategies3.4.3 GovernmentUNESCO, in collaboration with the Government of Namibia, is organized theUNESCO Southern Africa sub-Regional Forum on Artificial Intelligence, underthe theme ’Towards a sustainable development-oriented and ethical use of Arti-ficial Intelligence’ in Windhoek, Namibia, on 7-9 September 2022. This Forumwas the first sub-regional Forum on MI in Africa, following the first UNESCOForum on Artificial Intelligence in Africa in Benguérir, Morocco, in December2018 , which called for the organization of sub-regional Forums in Africa to fa-cilitate exchanges, elaboration of strategic frameworks and action plans in linewith unique sub-regional and national contexts, in view of an MI Strategy forAfrica.The Ministers and senior experts on MI from Southern African countriesconvened and deliberated on the development of a sustainable development-oriented and ethic use of MI in Southern Africa.The Windhoek Statement on Artificial Intelligence in Southern Africa madeseveral recommendations on MI in Africa.3.5 South Africa3.5.1 ResearchThe world around us is changing at a furious pace. It’s left many establishedbusinesses shaken, with executives questioning their organisations’ longevity. Toparticipate in a digital future, business transformation is critical and increasinglyurgent. A strategic approach is essential. To help businesses chart the wayforward, Accenture has partnered with the Gordon Institute of Business Science(GIBS) to provide insight into digital technologies and the future of business[238] .The Fourth Industrial Revolution has transformed modern society by usher-ing in the fusion of advances in robotics, the Internet of Things, genetic engineer-ing, quantum computing, and MI among others. MI brings a range of different43SA 2020-2023 Concrete ActionsResearch ✓ VulavulaIndustry ✓ DataProphetSMB ✓ Lelapa AI, Lesan,Xineoh, Clevva, Aer-obotics, The Gearsh,Credo, Akiba Digital,Bridgement, AfricAi, Neurozone, Congre-type, OpenbankingGovernment✓ AIISA: Artificial In-telligence Institute OfSouth AfricaTable 12: MI in South Africatechnologies and applications to interact with environments that comprise boththe relevant objects and the interaction rules and have the capacity to processinformation in a way that resembles intelligent behaviour. Similarly, artificialintelligence is also being used in the human resources management processes andfunctions in the public sector to map sequences to actions. The study exploresthe opportunities, challenges, and future prospects of integrating MI and PublicSector Human Resource Management in South Africa’s public sector. The studyin [239] examines the dynamics surrounding the adoption, implementation andoperationalisation of the 4IR in the management of human resources in the SApublic sector in this unfolding dispensation. Data was collected using the ex-tensive review of written records such as books, journal articles, book chaptersamong others which were selected for use in this study. Data was analysed usingcontent and thematic analysis techniques. The study established that ArtificialIntelligence is beneficial in the sense that it can improve public service deliveryin South Africa as the HRM personnel is enabled to focus more on the strategicareas of management by taking over routine tasks, and that it helps minimizebias in public service recruitment and selection. In contrast, research on po-tential challenges has revealed that combining MI and Public Sector HumanResource Management may pose a threat to white-collar jobs. This study maylead to practical applications of MI to support the HR functions of public sectorentities in SA. The public managers are better informed about the impediments,gaps and opportunities that may arise from using MI in managing human re-sources in SA’s public sector. This study contributes to the body of knowledgeas it unpacks and informs the dynamics associated with the implementation ofMI in managing human resources in public sector entities.Historically, South Africa has battled numerous chronic diseases such asCancer, Diabetes, and Tuberculosis. Even though significant efforts are madein the medical diagnostic industry to detect and treat these chronic diseases,44these efforts have fallen short due to their higher diagnostic costs, shortage ininfrastructure, equipment, and highly skilled technicians at the required time,resulting in reduced access to healthcare for patients. In recent years, the fieldof MI-based medical diagnosis has gained prominence because of its low cost,less infrastructure, equipment, and technician requirements. In addition, MI-based medical diagnosis reduces diagnostic time with a significantly high levelof accuracy. The work in [240] conducts a systematic literature review of 32collated MI articles. We present our findings, and scope the tools, techniques,and algorithms from a South African Context. The scope of this literaturereview involves (1) conducting an attribute analysis of literature that includesstudying disease, temporal, and spatial aspects of literature as well as stagesin developing MI-based medical diagnosis tool; (2) conducting a conceptualanalysis of literature that includes studying applications, algorithms, techniques,and performance measures related to different MI stages; and (3) scoping theinsights from the literature from a South African context that involves proposinga framework for developing MI-based medical diagnostic tool, hardware andsoftware requirements, and deployment strategies into underdeveloped medicaldiagnostic provinces of South Africa.The work in [270] focuses on adapting pretrained ASR models to low-resourceclinical speech using epistemic uncertainty-based data selection. In [271], theauthors propose FonMTL, a multitask learning approach for the Fon language.[272] introduces pretrained vision models for predicting high-risk breast can-cer stage. AfriNames, discussed in [273], addresses challenges in ASR modelsfor African names. [274] presents GFlowOut, a model incorporating dropoutwith generative flow networks. AfriQA, outlined in [275], focuses on cross-lingual open-retrieval question answering for African languages. AfriSpeech-200, as described in [276], introduces a pan-African accented speech datasetfor ASR. [277] investigates low-compute methods for low-resource African lan-guages. MasakhaNEWS, discussed in [278], involves news topic classificationfor African languages. The work in [279] focuses on MasakhaPOS, a part-of-speech tagging system for typologically diverse African languages. [280] inves-tigates language representation in multilingual models. PuoBERTa, introducedin [281], involves the training and evaluation of a curated language model forSetswana. MphayaNER, as discussed in [282], addresses named entity recogni-tion for Tshivenda. [283] explores fine-tuning multilingual pretrained Africanlanguage models. Unsupervised cross-lingual word embedding representationfor English-isiZulu is presented in [284]. Consultative engagement of stakehold-ers toward a roadmap for African language technologies is detailed in [285]. Thestudy in [286] evaluates the performance of large language models on Africanlanguages.3.5.2 IndustryBased in South Africa, DataProphet accelerates your smart factory journey withan integrated suite of AI-powered technology. These tools seamlessly integrateyour machine data with existing infrastructure, streaming high-quality data for45consistent, real-time access. The intuitive and secure platform for data visu-alization and KPI tracking across your plant facilitates a smooth transition tohigh-impact data-to-value decision-making. Prescriptive AI guides you in opti-mizing control plans and benchmarking manufacturing performance, enhancingefficiency and returns. DataProphet’s state-of-the-art MI tools seamlessly com-bine data orchestration, insightful contextualization, and digital-era continuousimprovement. Deep learning algorithms interpret complex process interactions,prescribing optimal recipes and precise steps for implementation. Plants wit-ness significant improvements in quality, production, and sustainability KPIs,achieving peak production efficiency.3.5.3 Small BusinessesLelapa AI: At the research level, A new venture called Lelapa AI, is trying to usemachine learning to create tools that specifically work for Africans. Vulavula,a new MI tool that Lelapa released, converts voice to text and detects namesof people and places in written text (which could be useful for summarizinga document or searching for someone online). It can currently identify fourlanguages spoken in South Africa - isiZulu, Afrikaans, Sesotho, and English - andthe team is working to include other languages from across Africa. The tool canbe used on its own or integrated into existing MI tools like online conversationalchatbots. Lesan: Lesan, a Berlin-based AI startup that is developing translationtools for Ethiopian languages. AJALA is a London-based startup that providesvoice automation for African languages. Lelapa AI and Lesan are just two of thestartups developing speech recognition tools for African languages. In February,Lelapa AI raised 2.5 million USD in seed funding, and the company plans forthe next funding round in 2025. But African entrepreneurs say they face majorhurdles, including lack of funding, limited access to investors, and difficulties intraining MI to learn diverse African languages.In South Africa, The Gearsh is a mobile app connecting artists with fans,providing a platform for bookings and learning. Credo is a Peer 2 Peer LendingPlatform, while Akiba Digital focuses on data and technology SAAS solutions.Bridgement, based in South Africa, helps small businesses optimize cash flowthrough its online platform offering instant advances on outstanding invoices.The AfricAi Project introduces DanAi, an MI-powered chatbot designed to rev-olutionize daily tasks with contextual knowledge for each African country. Neu-rozone, a neuroscience company in South Africa, optimizes brain/body systemsfor wellness and high performance, assembling expertise across various fields fora comprehensive Model of Brain Performance. Congretype Pty. Ltd primarilyfocuses on providing societal-based solutions in renewable energy, ICT for de-velopment, and Climate-Smart Agriculture. Trusource: A new secure, safe, andrevolutionary way for consumers to share financial information, allowing themto move, manage, and make more money with real-time payments and creditapprovals. All done on one platform powered by Trusource.Xineoh is a company based in Sandton, South Africa, that has created aplatform capable of predicting consumer behavior using machine intelligence.46For instance, VideoLlama is an implementation of Xineoh’s consumer behaviorprediction platform. It utilizes the platform to recommend movies and TV showsto users. The algorithm suggests films or TV shows to viewers one at a time.Based on user feedback on each suggestion being good, average, or bad, it learnsthe personal tastes of each viewer and recommends movies and shows they arein the mood for at that moment. Clevva, a South African technology company,specializes in decision navigation. It assists staff, customers, and digital workersthrough desired decision journeys, ensuring they consistently achieve the rightoutcomes in a contextually appropriate and compliant manner. Aerobotics,another South African company, uses machine intelligence and drones to helpfarmers manage their farms, trees, and fruits. Its technology tracks and assessesthe health of these crops, identifying when trees are sick, monitoring pests anddiseases, and conducting analyses for better yield management. The companyhas advanced its technology, providing farmers with independent and reliableyield estimates and harvest schedules by collecting and processing images oftrees and fruits from citrus producers early in the season. In turn, farmers canprepare their inventory, forecast demand, and ensure the best product qualityfor their customers.3.5.4 GovernmentThe Machine Intelligence Institute of Africa (MIIA) is an African non-profitorganization founded in 2016. MIIA aims to transform and help build an MI-powered Africa through a strong, innovative and collaborative Machine Intelli-gence, MI and Data Science community, consisting of individuals and key playersin the African Artificial Intelligence Ecosystem. MIIA’s growing network con-sists of stakeholders in the African MI Ecosystem, including thousands of mem-bers as well as key decision-makers in NGOs, NPOs, academia, businesses, andthe public sector. On 30 November 2022, The Department Of Communication& Digital Technologies launched the Artificial Intelligence Institute Of SouthAfrica (AIISA). On The Same Day, The University Of Johannesburg launchedthe First AI Hub Of AIISA. AIISA will focus on three sectors of the economy:the fourth industrial revolution in manufacturing, healthcare, agriculture andfood processing.The Centre for Artificial Intelligence Research (CAIR) is a South Africannational research network that conducts foundational, directed and appliedresearch into various aspects of Artificial Intelligence. CAIR has nodes atfive South African universities: the University of Cape Town, University ofKwaZulu-Natal, North-West University, University of Pretoria and StellenboschUniversity. CAIR was founded in November 2011 as a joint research centre be-tween the University of KwaZulu-Natal and the Council for Scientific and Indus-trial Research (CSIR). In 2015 CAIR expanded to four other South African uni-versities with the CSIR playing a coordinating role. CAIR is primarily funded bythe Department of Science and Technology (DST), as part of the implementationof South Africa’s ICT Research, Development and Innovation (RDI) Roadmap.Advances in ICT supported by the RDI Roadmap aim to guide South Africa to47a state of digital advantage that will strengthen economic competitiveness andenable an enhanced quality of life for all South Africans.4 Middle AfricaIn Central Africa, In Angola, there’s a focus on seismic inversion, forest fire de-tection, urban expansion monitoring, and bioavailable isoscapes. In Cameroon,MI is making significant strides in various sectors. In healthcare, MI is employedfor infant mortality rate analysis, tuberculosis detection through chest radiogra-phy interpretation, HIV management using medical imaging, and smartphone-based MI classifiers for cervical cancer screening. The country also sees agri-cultural innovation with MI-driven Agri-FinTech solutions and crop disease di-agnosis systems. Small businesses leverage MI for e-commerce enhancement,crop and soil monitoring, and skill development platforms. Additionally, there’sa government initiative establishing a MI training center to digitize the econ-omy. The Central African Republic focuses on IoT-based smart agriculture, MIpredicting electricity mix, and machine learning aiding in primate vocalizationclassification. The country is spearheading the development of a smart city, Re-naissance City, with MI consulting playing a pivotal role. The government leadsa crypto project, the Sango Initiative, introducing a digital-first economy backedby blockchain. The Sango platform further facilitates e-visas, digital companyregistration, and tokenization of natural resources. In Chad, MI is employed forfertility rate forecasting, conflict risk projections, hate speech detection in socialmedia, and predicting hotspots of food insecurity. The country’s entrepreneuriallandscape sees the emergence of an edtech solution, Genoskul, providing onlinetraining courses. Tech hub WenakLabs supports startups, offering incubation,mentorship, and financing opportunities. Chad is also actively working on aNational Cybersecurity Strategy to combat cyber threats. In the DemocraticRepublic of the Congo, MI is employed for gully erosion assessment using ma-chine learning methods like Random Forest and MaxEnt. Additionally, machinelearning and computer vision aid in property tax roll creation, addressing chal-lenges in resource-constrained environments. In the realm of small businesses,KivuGreen leverages MI to enhance climate resilience for farmers, showcasingthe positive impact on agriculture and livelihoods. Equatorial Guinea utilizesMI in researching sea level variability, demonstrating a critical application in cli-mate change mitigation. The country explores MI for economic diversification,recognizing its potential in reducing vulnerability to oil and gas price fluctu-ations. Gabon’s MI applications range from mapping land cover using cloudcomputing and machine learning to monitoring coastal erosion through convo-lutional neural networks. Deep learning frameworks are also employed for forestheight estimation, contributing to forest resource monitoring. The Republic ofthe Congo stands out with the establishment of the African Research Centre onMI. This pioneering center focuses on advancing research and digital technol-ogy, offering education and skills development to promote Africa’s integrationand inclusive economic growth. Sao Tomé and Principe showcases innovative48MI applications in social protection targeting through satellite data and valuechain analysis in the agricultural sector. These applications address challengesrelated to poverty mapping, program eligibility, and sustainable agriculturalpractices. Central Africa demonstrates a growing embrace of MI across varioussectors, showcasing both research advancements and practical applications thatcontribute to addressing local challenges and fostering economic development.Overall, MI is playing a transformative role across research, industry, smallbusinesses, and government initiatives in Central Africa.4.1 AngolaAngola 2020-2023 Concrete ActionsResearch ✓ cancerTable 13: MI in Angola4.1.1 Research[123] discusses Deep Learning Seismic Inversion as a case study from offshoreAngola. The deep learning seismic inversion results showed improved accuracyin identifying reservoir properties, surpassing traditional seismic inversion ap-proaches. [124] explores the use of GNSS reflectometry and machine learningto detect fire disturbances in forests in Angola. The research successfully ap-plies different machine learning techniques to identify burned areas, providingvaluable insights for monitoring forest disturbances. [125] focuses on Luanda,Angola. Their paper integrates geographical information systems, remote sens-ing, and machine learning to monitor urban expansion. The study evaluatesthe effectiveness of index-based classification and employs unsupervised machinelearning algorithms, revealing insights into the urban growth of Luanda.[126] presents a bioavailable strontium isoscape of Angola, crucial for un-derstanding the transatlantic slave trade. The study uses a machine learningframework to create the first bioavailable 87Sr/86Sr map for Angola, aiding inidentifying the geographic origin of enslaved individuals and contributing to ar-chaeological and forensic studies. In [127] Zolana R Joao’s research focuses ona ”Road Construction Assessment Model (RC-AM)” designed to prevent con-tract overbilling in Angola. The study explores the use of satellite imagery andmachine learning to identify road layers and measure road lengths, providingtriggers for road inspections and addressing the issue of poor road quality. [128]Investigates stored maize in Angola, this study identifies insects and fungi asso-ciated with maize under various storage conditions. The findings contribute toimproving food security in Angola by understanding the pests affecting storedmaize in different provinces, offering insights into pest management strategies.In [129] Geraldo AR Ramos and Lateef Akanji apply MI for technical screen-ing of enhanced oil recovery (EOR) methods. The study utilizes a five-layered49feedforward backpropagation algorithm, incorporating fuzzy logic reasoning andneural networks to screen and predict suitable EOR techniques for oilfields.[130]. In this work, Geraldo AR Ramos and Lateef Akanji employ a neuro-fuzzy simulation study to screen candidate reservoirs for enhanced oil recoveryprojects in Angolan oilfields. The study combines fuzzy logic and neural net-work techniques to assess the suitability of EOR techniques based on rock andfluid data. [131] presents a case study on ”3D Water Saturation EstimationUsing AVO Inversion Output and Machine Neural Network” in offshore Angola.The study utilizes high-resolution seismic data and neural network techniquesto estimate water saturation in the reservoir, providing insights for hydrocarbonprospection and development.4.2 CameroonCameroon 2020-2023 Concrete ActionsResearch ✓ HealthcareSMB ✓ Agrix Tech, Com-paroshop, eFarm,KMER MI, DAS-TUDYGovernment✓ AI training centerTable 14: MI in Cameroon4.2.1 ResearchIn a recent study [87], researchers utilized artificial Neural Networks to analyzethe infant mortality rate in Cameroon. The data, spanning from 1960 to 2020,with projections reaching 2030, demonstrated the model’s stability in forecast-ing. According to the model (12, 12, 1), the predicted Infant Mortality Rate(IMR) is approximately 48 per 1000 live births annually. This suggests a needfor government action, including strengthening primary healthcare, enhancingimmunization coverage, and providing training for health workers to reduce in-fant mortality.In a study [88], Deep Learning neural networks were employed to interpretchest radiography (CXR) for screening and triaging pulmonary tuberculosis(TB). This retrospective evaluation assessed three DL systems (CAD4TB, Lu-nit INSIGHT, and qXR) for detecting TB-associated abnormalities in chestradiographs from outpatients in Nepal and Cameroon. All 1196 individualsunderwent Xpert MTB/RIF assay and CXR readings by radiologists and DLsystems, with Xpert as the reference standard. DL systems showed higher speci-ficities compared to radiologists, potentially reducing Xpert MTB/RIF tests by66% while maintaining sensitivity at 95% or better. However, performance vari-ations across sites emphasize the importance of selecting scores based on the50screened population. DL systems are valuable for TB programs with limitedhuman resources and available automated technologyIn a study [89], the focus is on combating the spread of HIV using MedicalImaging. With the global challenge posed by HIV, there’s a hot debate on ap-plying MI to manage the disease. Over 6,000 people are newly diagnosed withHIV annually in the United States, where an estimated 1.2 million have the dis-ease. Ongoing medical care for HIV-positive individuals is crucial, and recentstudies suggest that MI can enhance the speed and accuracy of HIV detection.For example, a Northwestern University study used a deep learning algorithmto identify HIV-positive individuals with 94.2% accuracy, outperforming man-ual diagnosis. MI, especially MI, is also improving patient-specific treatmentstrategies, as demonstrated by a University of California, San Francisco study.Tailored treatment regimens based on machine learning algorithms proved moreaccurate than conventional methods. MI is further utilized to monitor the virus’sdevelopment in HIV-positive individuals, as seen in a Stanford University study.With faster and more accurate diagnoses, personalized treatment plans, andenhanced virus monitoring, MI has the potential to revolutionize HIV control,offering more effective and efficient treatment for patients.In a study [90], the focus is on a protocol for a two-site clinical trial vali-dating a smartphone-based MI classifier for identifying cervical precancer andcancer in HPV-positive women in Cameroon. Cervical cancer remains a signifi-cant public health challenge in low- and middle-income countries (LMICs) dueto financial and logistical issues. The WHO recommends HPV testing as theprimary screening method in LMICs, followed by visual inspection with aceticacid (VIA) and treatment. However, VIA is subjective and dependent on thehealthcare provider’s experience. To enhance accuracy, our study aims to assessthe performance of a smartphone-based Automated VIA Classifier (AVC) usingMI to distinguish precancerous and cancerous lesions from normal cervical tis-sue. The AVC study is nested in the ongoing cervical cancer screening program”3T-study” (Test, Triage, and Treat), involving HPV self-sampling, VIA triage,and treatment if needed. After applying acetic acid on the cervix, precancerousand cancerous cells whiten more rapidly than non-cancerous ones, and theirwhiteness persists stronger over time.4.2.2 Small BusinessesAgrix Tech, headquartered in Yaoundé, Cameroon, is an Agri-FinTech companydedicated to assisting small-scale farmers in transitioning from subsistence farm-ing to commercial ventures, aiming to maximize their profits. The company pro-vides farmers with a comprehensive package, including financing, farm inputs,expert advice, insurance, and, whenever possible, access to markets. Leverag-ing machine learning and satellite data, Agrix Tech facilitates improved creditdecisions, while automated operations ensure cost-effectiveness and scalability.AllGreen startup introduces ROSALIE, an MI system designed to diagnosecrop diseases in real-time with accuracy. Beyond diagnosis, ROSALIE also offersnotifications to guide owners on adopting effective practices for eradicating these51diseases, contributing to improved crop health and yields.Comparoshop operates as a SaaS toolkit utilizing MI to enhance e-commercein Africa. The platform equips web merchants with solutions, enabling themto set up a sales site in just 5 minutes, automate product catalog updates,enhance user experience, increase shopping cart conversions, and stay updatedon real-time market trends and competition.eFarm, a web and mobile platform for agricultural products, incorporates MIto optimize crop and soil monitoring, enhancing agriculture’s efficiency and effi-cacy. This innovative approach, developed during TechStars Start-up Weekend,aims to significantly increase productivity, playing a crucial role in the globalfight against hunger and facilitating access to markets for farmers worldwide.DASTUDY, an MI-type platform, is a local innovation developed by Cameroo-nian entrepreneur David Kenfack. This platform enables users to acquire newskills, share content, receive answers based on local socio-cultural realities, andseek assistance with daily tasks such as administrative writing, analysis, andtranslation. DASTUDY serves as a knowledge-sharing platform between learn-ers and professionals, facilitating the exchange of documents, exercises, andsoftware in academic, professional, and cultural fields. Additionally, it featuresa virtual assistant named TSAF, powered by Generative MI, designed by thisCameroonian startup. The platform addresses the need to enhance local skillsand provide visibility to content producers in the local community.KMER AI is a platform dedicated to promoting MI and showcasing inno-vative technologies and research in Cameroon. Our goal is to educate studentson the fundamentals of MI and provide a space for Cameroonian researchers,students, and industries to share their work and interests in the field. We alsoaim to create an opportunity for the relatively few Cameroonians involved inthe field to collaborate, network, and communicate, motivating and encouragingothers to join this journey and address challenges in our own context. By unit-ing theoretical and applied machine learning techniques in cutting-edge areassuch as health, agriculture, climate change, marketing, education, transporta-tion, linguistics, and art, we aspire to generate impactful ideas. These ideas willinspire students and companies to harness the power of MI, contributing to theacceleration of Cameroon’s emergence.4.2.3 GovernmentIn 2019, Cameroon’s first MI training center has been established through apartnership between the state-owned telecommunications operator Camtel andthe university of Yaounde I. The training facility will be hosted by the Uni-versity’s National Advanced School of Engineering in Yaounde, which alreadyhouses a high-tech 3D printing center. The project, valued at FCFA 1.3 bil-lion, covers the construction of learning infrastructure, procurement of trainingequipment, training of trainers, curriculum design, and learner scholarships,aiming to be fully operational within two and a half years. As part of the coun-try’s digitization strategy, there’s a target to multiply the number of direct andindirect jobs in ICT by 50. The establishment of the training resource aligns52with the government’s vision to digitize the economy and develop relevant skillsets. The center plans to run research and development postgraduate programswith market-related portfolios, intending to improve the employability of job-seekers through transformative education and training systems. The goal is tomeet the knowledge, competencies, skills, research, innovation, and creativityneeds required to nurture the future of the MI sector. The unnamed managerstated that they envision training one hundred individuals initially, with twenty-five percent of pioneer students benefiting from fully-funded scholarships offeredby the telecommunications operator and its partners, as well as the universitythrough Polytech. The project aims to position Cameroon among the leadersin MI training on the continent.4.3 Central African RepublicCAR 2020-2023 Concrete ActionsResearch ✓ electricitySMB ✓InformalEconomy✓Government✓ Web3 Sango, NationalBitcoin TreasuryTable 15: MI in Central African Republic4.3.1 ResearchDesertification poses a threat to the Central African Republic’s economy heavilyreliant on agriculture, especially in arid zones with low rainfall. To address this,[91] introduces a smart agriculture solution using the Internet of Things. Thesystem, based on a Raspberry Pi 3 B+, employs temperature and soil moisturesensors, a submersible water pump, and a relay. TV Whitespace (TVWS) isthe inactive or unused space found between channels actively used in UHF andVHF spectrum. TVWS frequency spans from 470 MHz - 790 MHz. Using aTVWS network, farmers access a web interface on their smartphones to monitorfield temperature and soil moisture, remotely activating or deactivating irriga-tion. This solution aims to enhance economic resilience and promote sustainableagriculture in arid regions. In [92], a machine-learning model is developed topredict Africa’s electricity mix based on planned power plants and their successlikelihood. This addresses the limitations of energy scenarios, offering insightsinto the risks associated with planned power-generation projects. By utilizinga large dataset and country-level characteristics, the model accurately predictsproject outcomes. Contrary to some rapid transition scenarios, the study sug-gests that non-hydro renewables in Africa’s electricity generation may remainbelow 10% in 2030, highlighting potential carbon lock-in risks. [93] explores53the automatic classification of primate vocalizations in Central Africa usingdeep learning. The study compares various neural network architectures andemploys data augmentation techniques to address the small training dataset.The best model, a standard 10-layer CNN, achieves high accuracy in classifyingprimate species. The results showcase the effectiveness of augmentations andtraining tricks, providing insights into improving the classification of primatevocalizations. In the context of 12 years of politico-military crises in the Cen-tral African Republic, [94] proposes algorithms for remediating lost school anduniversity years. These algorithms are implemented using an API FrameworkRasa chatbot, considering the social imbalance in education caused by conflictsand worsened by the Covid-19 pandemic. The chatbot offers guidance on nor-mal, professional, and specific training, considering the learner’s province andthe Covid-19-induced social distancing measures. [95] focuses on the applica-tion of machine learning for diagnosing 18 common pediatric diseases in theCentral African Republic. The study addresses the healthcare challenges in theregion, developing a diagnostic system based on decision tree, random forest,and neural network models. Results indicate high diagnostic accuracy, showcas-ing the potential of MI in improving healthcare access and quality in developingcountries.4.3.2 Small BusinessesRenaissance City is set to become the pioneering smart city in the heart ofAfrica, situated in the Central African Republic, ushering in new business andcommunity prospects. WebRobot, the official MI and Big-Data consulting com-pany for this ambitious endeavor, will play a key role in shaping a sustainable,innovative, technological, and economic hub. In collaboration with the Mis-tahou Financial Group, this project aims to propel Renaissance City to theforefront of smart city development in Africa.4.3.3 GovernmentThe Sango Initiative, initiated by the Central African Republic National As-sembly and endorsed by the President, is a crypto project aimed at fosteringa digital-first, blockchain-based economy. Project Sango, launched on July 15,2022, introduces the Sango coin, representing the country’s move towards a dig-ital future. This national digital currency, backed by Bitcoin, forms the core ofa new digital monetary system, including a ”Digital National Bank” and ”Na-tional Bitcoin Treasury.” The government’s collaboration with the private sectormirrors a public-private partnership. Expanding the Sango blockchain project,the Central African Republic has introduced tokenization of its land and natu-ral resources as of July 2023. The E-visa, facilitated through the Web3 Sangoplatform, is a digitally issued visa streamlining entry into the Central AfricanRepublic for specific purposes. Unlike traditional visas, the e-visa application,payment, and approval occur digitally with the Sango platform’s integratedwallet feature. Digital company registration involves the official incorporation54of businesses through the Web3 Sango platform powered by blockchain. Thisprocess, payable in cryptocurrency, allows submitting necessary documents togovernment authorities for establishing a legal entity for business in the CentralAfrican Republic. In blockchain technology, tokenization represents the digi-tal representation of ownership and rights over the Central African Republic’snatural resources on the Sango blockchain. This approach enhances manage-ment, monitoring, and international investment opportunities in the country’sresources. The Sango tokenization platform offers various agricultural and forestplots across the country, enabling investors to concession and engage in agricul-ture or forest exploitation. Investors can further tokenize their concessions toattract additional investment.4.4 Chad4.4.1 ResearchChad 2020-2023 Concrete ActionsResearch ✓ lingua francaSMB ✓ Genoskul, WenakLabs,DaTchad, ZereSoftInformalEconomy✓Government✓ National CybersecurityStrategyTable 16: MI in Chad[97] employs machine learning algorithms to analyze remote sensing andground-truth Lake Chad’s level data. The study addresses environmental chal-lenges faced by Lake Chad due to climate change and anthropogenic activi-ties. Results reveal associations between climate variables, remote sensing, andground-truth lake levels. Random Forest Regression outperforms other models,emphasizing soil temperature’s role in remote sensing lake level fluctuations.The study provides insights for integrated water management in the Lake Chadbasin, considering climate change, vulnerability, human activities, and waterbalance. [96] forecasts the Total Fertility Rate in Chad using a machine learn-ing approach. The study applies an Artificial Neural Network to analyze TFRdata from 1960 to 2018, with out-of-sample forecasting for 2019-2030. Modelevaluation criteria indicate its stability, predicting an increase in annual totalfertility rates in Chad. Recommendations include boosting demand for fam-ily planning services, enhancing accessibility to sexual and reproductive healthservices, and empowering women through education, labor participation, andpromoting women’s rights.[98] projects armed conflict risk in Africa towards 2050 using a machinelearning approach. Utilizing the CoPro ML framework, the study explores sub-55national conflict risk for different socio-economic and climate scenarios. Resultsalign with socio-economic storylines, projecting conflict intensification in severescenarios. The study identifies hydro-climatic indicators’ limited but contex-tual role in conflict drivers. Challenges include inconsistent data availability,but ML models present a viable approach for armed conflict risk projections,informing climate security policy-making. [99] Forecasts conflict in Africa us-ing automated machine learning systems. The ViEWS competition focuses onpredicting changes in the level of state-based violence for the next six monthsat the PRIO-GRID and country level. The study explores combinations of au-toML systems and limited datasets, emphasizing the endogenous nature of con-flict. Key findings include the improved predictive performance of autoML andthe superiority of the Dynamics model. The Dynamics model, utilizing limiteddata related to state-based violence and its spatial-temporal structure, won theViEWS competition for ”predictive accuracy” at the PGM level. [100] PredictsHotspots of Food Insecurity in Chad. The paper addresses the persistent is-sue of food insecurity, examining factors such as poverty, conflicts, and climatecontributing to the lack of access to nutritious food. By forecasting hotspotregions for food insecurity, the study aims to provide program managers withlead time to organize and coordinate efforts. Chad, with high hunger rates andpoverty exacerbated by environmental degradation, desertification, and conflict,is the focus. The paper classifies food insecurity based on the Cadre Harmoniséphases, offering policy recommendations. [101] Examines Measurements anddeterminants of extreme multidimensional energy poverty using machine learn-ing. The study calculates the depth, intensity, and degrees of energy poverty indeveloping countries, revealing widespread ’severe’ energy poverty across multi-ple dimensions. Machine learning identifies the most influential socioeconomicdeterminants of extreme multidimensional energy poverty, including householdwealth, house size and ownership, marital status of the main breadwinner, andresidence of the main breadwinner. The findings have policy significance foraddressing severe energy poverty through incentives, resource allocation, andspecial assistance. [102] Studies Delineation of Groundwater Potential Zonesin the Eastern Lake Chad Basin using ensemble tree supervised classificationmethods. The paper employs machine learning to map groundwater potential incrystalline domains. Twenty classifiers are trained on 488 boreholes and exca-vated wells, with random forest and extra trees classifiers outperforming others.Relevant explanatory variables include fracture density, slope, SAR coherence,topographic wetness index, basement depth, distance to channels, and slopeaspect. The study emphasizes the importance of using a large number of clas-sification algorithms, the impact of performance metrics on variable relevance,and the contribution of seasonal variations in satellite images to groundwaterpotential mapping. The work in [103, 104] examines the Detection of HateSpeech Texts Using Machine Learning Algorithm. The article focuses on identi-fying hate speech in social media, particularly in ”lingua franca,” a mix of localChadian and French languages. The dataset consists of 14,000 comments frompopular Facebook pages, categorized as hate, offense, insult, or neutral. Natu-ral Language Processing techniques clean the data, and three word embedding56methods (Word2Vec, Doc2Vec, Fasttext) are applied. Four machine learningmethods (LR, SVM, RF, KNN) classify different categories, with the FastText-SVM combination achieving 95.4% accuracy in predicting comments containinginsults. [105] examines a Machine Learning-Based Model of Boko Haram. Thebook presents a computational modeling effort to understand Boko Haram’s be-havior, gathering data from 2009 to 2016. Predictive models generate forecastsof Boko Haram attacks every month, allowing real-world testing of accuracy.The book introduces Temporal Probabilistic (TP) rules to explain predictionsand enhance understanding of Boko Haram’s behaviors for counter-terrorismanalysts, law enforcement, policymakers, and diplomats.4.4.2 Small BusinessesGenoskul is an edtech solution developed by a Chadian start-up, providing ac-cess to online training courses and tutors. It features a smart assistant todeliver relevant answers to users’ queries. Through its Android app, users canregister using an email or phone number, gaining access to services like virtualclassrooms for discussions with fellow learners. These virtual rooms connectlearners from diverse backgrounds for intellectual exchange, supervised by qual-ified teachers to prepare effectively for national and international secondaryand higher education exams and competitions. Genoskul offers courses span-ning various professions, from loincloth shoe making and shea butter processingto rabbit breeding, public management, sustainable development, and civic ac-tion. In support of its growth, Genoskul secured CFAF 5 million (approximately$8,149) in funding and received backing from Chad Innovation, an incubatorthat provided the start-up with a stand at Gitex Africa 2023 in Marrakech,Morocco.In the emerging entrepreneurial landscape of Chad, tech hub WenakLabsstands at the forefront of innovation and creativity. Located in N’Djamena, thehub offers incubation, mentorship, and financing opportunities to startups, fa-cilitating the transformation of ideas into thriving companies. Founded in 2014through collaboration between Chadian bloggers, the French platform Mon-doblog, local network JerryClan Tchad, and tech entrepreneur Abdelsalam Safi,who currently serves as the CEO of WenakLabs. Over the years, the hub hasclosely partnered with local and international entities such as AfriLabs, Oxfam,the French Institute, Moov Africa, Sahel Innov, and UNICEF. This collabo-ration aims to provide resources and tools supporting Chadian entrepreneursin launching and growing their businesses. WenakLabs encourages youth en-trepreneurship by offering a fab lab and a media lab. The fab lab serves asan open-access digital production space, providing digitally controlled machinesto the public for designing and creating physical objects collaboratively. Itis open to young individuals seeking to acquire skills for quickly transform-ing their ideas into physical prototypes. The media lab provides an innovativecommunication environment for incubatees and project leaders, utilizing newtechnologies to process, visualize, and share information. It serves as a plat-form to disseminate reliable and engaging information, inspiring communities57to take action. With approximately 120 startups incubated and 70 projectsdeveloped in Chad, WenakLabs boasts achievements such as ZereSoft, a plat-form modernizing agriculture and the rural world with 2.0 tools, DaTchad, adata journalism agency project, and Nomad Learning, an SMS-based learningplatform. The incubator actively organizes programs and events, including theStartup Weekend N’Djamena and DENE MAGIC, a 2022 initiative aiming toprovide digital skills to women for better access to decent jobs. WenakLabs alsoprovides advice, training, and coaching to socioeconomic and digital develop-ment actors in the country.4.4.3 GovernmentChad is on the path to establishing a National Cybersecurity Strategy. TheMinistry of Telecommunications and Digital Economy, along with the NationalAgency for Computer Security and Electronic Certification, launched the devel-opment of this strategy on December 14, with the Telecommunications Ministerin attendance. Developed in partnership with the International Telecommuni-cation Union, the future National Cybersecurity Strategy aims to find means tobetter combat all forms of cyber attacks. ”It is important to assess the stakesrelated to cybersecurity to define and prioritize responses to establish a strat-egy capable of providing greater digital security to all structures. To strengthenthe regulations, the government has decided to make significant progress in im-plementing the National Cybersecurity Strategy, which has lagged behind forsome years. In 2019, a gathering involving participants from 32 national andregional institutions took place in the country. One of the resolutions fromthe discussions was to accelerate the process of developing the national cyber-security strategy in Chad. In February 2022, Chad also hosted cybersecurityexperts from various countries and the sub-region to discuss issues related toevaluation methodology, strategic cybersecurity policy, online commerce, bank-ing, legal and regulatory frameworks, and technology standards. In December2022, Chad accelerated its efforts to strengthen cybersecurity. On December 5,two bills were adopted to enhance the country’s cybersecurity: the first rati-fies Ordinance No. 007/PCMT/2022 of August 31, 2022, regarding cybercrimeand cyber defense, and the second ratifies Ordinance No. 008/PCMT/2022 ofAugust 31, 2022, regarding cybersecurity.4.5 Democratic Republic of the Congo4.5.1 Research[106] considers four machine learning methods to examine gully erosion in Demo-cratic Republic of the Congo. Soil erosion by gullying causes severe soil degra-dation, resulting in profound socio-economic and environmental damages intropical and subtropical regions. To mitigate these adverse effects and ensuresustainable natural resource management, preventing gullies is imperative. Ef-fective gully management strategies begin with devising appropriate assessment58DRC 2020-2023 Concrete ActionsResearch ✓SMB ✓ KivuGreenInformalEconomy✓Government✓ e-health, e-learningTable 17: MI in Democratic Republic of the Congotools and identifying driving factors and control measures. Machine learningmethods play a crucial role in identifying these driving factors for implement-ing site-specific control measures. Their study aimed to assess the effectivenessof four machine learning methods (Random Forest (RF), Maximum Entropy(MaxEnt), Artificial Neural Network, and Boosted Regression Tree (BRT)) inidentifying gully-driving factors and predicting gully erosion susceptibility inthe Luzinzi watershed, Walungu territory, eastern Democratic Republic of theCongo. Gullies were initially identified through field surveys and digitized usinga high-resolution image from Google Earth. Of the 270 identified gullies, 70%(189) were randomly selected for training the machine learning methods withtopographical, hydrological, and environmental factors. The remaining 30% (81gullies) were used for testing the methods. They have used the area underthe receiver operating characteristic (AUROC) method. Results showed thatRF and MaxEnt algorithms outperformed other methods, with RF (AUROC= 0.82) and MaxEnt (AUROC = 0.804) exhibiting higher prediction accuraciesthan BRT (AUROC = 0.69) and ANN (AUROC = 0.55). TSS results indicatedthat RF and MaxEnt were the best methods in predicting gully susceptibility inLuzinzi watershed. Factors like Digital Elevation Model, Normalized DifferenceWater Index, Normalized Difference Vegetation Index, slope, distance to roads,distance to rivers, and Stream Power Index played key roles in gully occurrence.Considering these factors is crucial for policymakers to develop strategies toreduce the risk of gully occurrence and related consequences at the watershedscale in eastern DRC.Developing countries often face financial constraints in providing publicgoods. Property taxation is seen as a promising local revenue source due toits efficiency, ability to capture real estate value growth, and potential for pro-gressivity. However, ineffective property tax collection is common in many low-income countries, often due to missing or incomplete property tax rolls. In alarge Congolese city, the work in [107] employs machine learning and computervision models to build a property tax roll. Training the algorithm on 1,654 ran-domly selected properties assessed during government land surveyors’ in-personvisits, along with property characteristics from administrative data or extractedfrom photographs, we achieve promising results. The best machine learning al-gorithm, trained on administrative data, achieves a cross-validated R2 of 60%,with 22% of predicted values within 20% of the target value. Computer vision59algorithms, relying on property picture features, perform less effectively, withonly 9% of predicted values within 20% of the target value for the best algorithm.These findings suggest that even in contexts with limited property information,simple machine learning methods can assist in constructing a property tax roll,particularly when the government can only collect a small number of propertyvalues through in-person visits.4.5.2 Small BusinessesIn North Kivu, a province of the Democratic Republic of the Congo, manypeople earn their livelihood as small-scale farmers. However, climate change ismaking farming increasingly challenging due to unpredictable weather patterns.This not only threatens their income but also their food security. KivuGreen,a youth-led enterprise, is enhancing the climate resilience of small-scale farmersin the Democratic Republic of the Congo through a mobile service that providesreal-time forecasts and climate-smart agricultural advice. Building the adaptivecapacity of these farmers faced challenges, such as limited internet connectivityin rural areas and the prevalence of older mobile phones. To overcome thesehurdles, KivuGreen’s service is made accessible via SMS, eliminating the needfor expensive technology or an internet connection. The impact has been sig-nificant, with small-scale farmers increasing their agricultural yield by 40% andtheir income by 30%. In addition to enhancing food security, the increasedincome allows farmers to invest in their children’s education, healthcare, sani-tary facilities, and non-polluting energy sources. By boosting the agriculturalyield and income of small farmers, KivuGreen contributes to creating more jobopportunities for young people in North Kivu.4.5.3 GovernmentIn 2022, the government of Democratic Republic of Congo organized a forumunder the theme ”Artificial intelligence (AI), myth or reality”. According tothe authorities of this Central African country, the capital Kinshasa populatedby 17 million souls will soon be a smart city thanks to the implementation ofthe “Smart City” project launched in 2019.In low-income countries (such as the Democratic Republic of the Congo) aspecific challenge for public health professionals is how to accurately distinguishmalaria symptoms from other febrile illnesses with similar characteristics. Inremote rural areas, microscopic diagnosis is slow, inaccurate and done in non-specialized laboratories; lab technicians have difficulty reading smear resultsto rapidly determine the species and stage of plasmodium, putting healthcareprofessionals and their patients in these areas at a particular disadvantage.Through a web connection the system will support physicians and healthcareprofessionals in any location [in the country] rapidly identify malaria type andseverity for individual patients and so prescribe the optimal treatmentA Prototype web-based platform and tools in the Democratic Republic ofthe Congo to connect rural health care professionals with malaria experts. The60e-health platform will also have a growing library of e-learning resources onmalaria – including articles, reports, videos, documentation and quizzes – thathealthcare professionals can consult to deepen their knowledge of diagnosis andtreatment.4.6 Equatorial GuineaEquatorialGuinea2020-2023 Concrete ActionsResearch ✓SMB ✓InformalEconomy✓Government✓Table 18: MI in Equatorial Guinea4.6.1 Research[108] investigates sea level variability and predictions using artificial neural net-works and other machine learning techniques in the Gulf of Guinea. The risingsea level due to climate change poses a critical threat, particularly affecting vul-nerable low-lying coastal areas such as the Gulf of Guinea (GoG). This impactnecessitates precise sea level prediction models to guide planning and mitiga-tion efforts for safeguarding coastal communities and ecosystems. The studypresents a comprehensive analysis of mean sea level anomaly (MSLA) trends inthe GoG between 1993 and 2020, covering three distinct periods (1993–2002,2003–2012, and 2013–2020). It investigates connections between interannualsea level variability and large-scale oceanic and atmospheric forcings. Addi-tionally, the performance of artificial neural networks (LSTM and MLPR) andother machine learning techniques (MLR, GBM, and RFR) is evaluated to opti-mize sea level predictions. The findings reveal a consistent rise in MSLA lineartrends across the basin, particularly pronounced in the north, with a total lineartrend of 88 mm/year over the entire period. The highest decadal trend (38.7mm/year) emerged during 2013–2020, and the most substantial percentage in-crement (100%) occurred in 2003–2012. Spatial variation in decadal sea-leveltrends was influenced by subbasin physical forcings. Strong interannual signalsin the spatial sea level distribution were identified, linked to large-scale oceanicand atmospheric phenomena. Seasonal variations in sea level trends are at-tributed to seasonal changes in the forcing factors. Model evaluation indicatesRFR and GBR as accurate methods, reproducing interannual sea level patternswith 97% and 96% accuracy, respectively. These findings contribute essentialinsights for effective coastal management and climate adaptation strategies in61the GoG. [109] explores the spatial ecology and conservation of leatherback tur-tles (Dermochelys coriacea) nesting in Bioko, Equatorial Guinea. Bioko Island(Equatorial Guinea) hosts essential nesting habitat for leatherback sea turtles,with the main nesting beaches found on the island’s southern end. Nest moni-toring and protection have been ongoing for more than two decades, althoughdistribution and habitat range at sea remain to be determined. This study usessatellite telemetry to describe the movements of female leatherback turtles (n= 10) during and following the breeding season, tracking them to presumed off-shore foraging habitats in the south Atlantic Ocean. Leatherback turtles spent100% of their time during the breeding period within the Exclusive EconomicZone of Equatorial Guinea, with a core distribution focused on the south ofBioko Island extending up to 10 km from the coast. During this period, turtlesspent less than 10% of time within the existing protected area. Extending theborder of this area by 3 km offshore would lead to a greater than threefoldincrease in coverage of turtle distribution (29.8 ± 19.0% of time), while an ex-pansion to 15 km offshore would provide spatial coverage for more than 50% oftracking time. Post-nesting movements traversed the territorial waters of SaoTomé and Principe (6.4% of tracking time), Brazil (0.85%), Ascension (1.8%),and Saint Helena (0.75%). The majority (70%) of tracking time was spent inareas beyond national jurisdiction (i.e. the High Seas). This study reveals thatconservation benefits could be achieved by expanding existing protected areasstretching from the Bioko coastal zone, and suggests shared migratory routesand foraging space between the Bioko population and other leatherback turtlerookeries in this region.[110] explores knowledge about Fang Traditional Medicine: an informalhealth-seeking behavior for medical or cultural afflictions in Equatorial Guinea.This study delves into a range of informal health-seeking behaviors, includingthe use of Fang Traditional Medicine (FTM) for medical or cultural afflictions inEquatorial Guinea (EQ). The research covers therapeutic methods, health prob-lems addressed, the learning process, traditional medicine user profiles, and thesocial images of Fang Traditional Healers (FTHs). Ethnography was employedas a qualitative strategy using emic-etic approaches. Semi-structured interviewswere conducted with 45 individuals, including 6 community leaders, 19 tribalelders, 7 healthcare professionals, 11 FTHs, and 2 relatives of traditional healersin 5 districts of EQ. FTM offers a cure for malaria and treatments for reproduc-tive health issues, bone fractures, and cultural illnesses. Several methods used tolearn FTM are based on empirical observation, without the need for traditionalschooling. For example, watching a family member or the spirits/ancestorscan reveal healing knowledge. Materials from forests, including tree barks andplants, and rituals are used to keep Fang populations healthy. In addition, tworituals known as ’osuin’ and ’etoak’ (infusions of tree barks with the blood ofsacrificed animals) are the most commonly used treatments. Elders and womenare the most active consumers of FTM, playing a relevant role in curing med-ical and cultural afflictions in Fang communities. The informal health-seekingbehavior among the Fang community is conditioned by the explanation modelof illness.624.6.2 GovernmentEquatorial Guinea’s economic landscape, predominantly reliant on the oil andgas sector, presents a vulnerability due to its susceptibility to fluctuating globalprices and the finite nature of these resources. Embracing MI technologiesoffers a strategic avenue for economic diversification. By investing in researchand development in MI, Equatorial Guinea can foster innovation across varioussectors.4.7 GabonGabon 2020-2023 Concrete ActionsResearch ✓ Grand LibrevilleIndustry ✓ Oil and GasTable 19: MI in Gabon4.7.1 Research[111] examines cloud computing and machine learning in support of country-level land cover and ecosystem extent mapping in Liberia and Gabon. Liberiaand Gabon joined the Gaborone Declaration for Sustainability in Africa, estab-lished in 2012, with the goal of incorporating the value of nature into nationaldecision-making by estimating the multiple services obtained from ecosystemsusing the natural capital accounting framework. In this study, we produced 30-m resolution, 10-class land cover maps for the 2015 epoch for Liberia and Gabonusing the Google Earth Engine cloud platform to support the ongoing naturalcapital accounting efforts in these nations. We propose an integrated methodof pixel-based classification using Landsat 8 data, the Random Forest classifier,and ancillary data to produce high-quality land cover products for a broad rangeof applications, including natural capital accounting. Our approach focuses ona pre-classification filtering (Masking Phase) based on spectral signature andancillary data to reduce the number of pixels prone to be misclassified, therebyincreasing the quality of the final product. The proposed approach yields anoverall accuracy of 83% and 81% for Liberia and Gabon, respectively, outper-forming prior land cover products for these countries in both thematic contentand accuracy. Their approach is replicable and was able to produce high-qualityland cover products to fill an observational gap in up-to-date land cover dataat the national scale for Liberia and Gabon. [112] studies the Grand Libre-ville, Gabon coastline using machine learning and convolutional neural networkdetection and automatic extraction methods. Coastal erosion, worsened by cli-mate change and natural occurrences like droughts and marine flooding, is amajor problem. Countries situated along coastlines face significant challengesin preserving their land and protecting their people and assets. To mitigate63the damage caused by sea encroachment on land, effective monitoring toolsand methods are required. This study uses Object-oriented Analysis (OBIA),Pixel-Oriented Analysis (PBIA), and Convolutional Neural Network methodsto automatically detect and extract the Greater Libreville coastline based onPléiades very high-resolution satellite images dating from 2022. Three test areaswere chosen and extracted, showing competitive Overall Accuracy values withthe OBIA method and the CNN model. However, the OBIA method using theRandom Forest algorithm achieved the highest accuracy rates, reaching 95%,90%, and 80% for the three test areas, respectively. [113] utilizes a deep learn-ing framework for the estimation of forest height from bistatic TanDEM-X data.Up-to-date canopy height model estimates are crucial for forest resource moni-toring and disturbance analysis. This study presents a deep learning approachfor the regression of forest height from TanDEM-X bistatic interferometric syn-thetic aperture radar (InSAR) data. The proposed fully convolutional neuralnetwork framework is trained using reference CHM measurements derived fromthe LiDAR LVIS airborne sensor from NASA, acquired during the joint NASA-ESA 2016 AfriSAR campaign over five sites in Gabon, Africa. The DL modelachieves an overall performance of 1.46-m mean error, 4.2-m mean absoluteerror, and 15.06% mean absolute percentage error when tested on all consid-ered sites. Additionally, a spatial transfer analysis provides insights into thegeneralization capability of the network when trained and tested on datasetsacquired over different locations and types of tropical vegetation. The resultsare promising and align with state-of-the-art methods based on both physical-based modeling and data-driven approaches, with the advantage of requiringonly one single TanDEM-X acquisition at inference time.4.7.2 IndustryNeural networks have been used in the oil and gas industry in [115, 116, 117,118, 119, 120].4.8 Republic of the CongoCongo 2020-2023 Concrete ActionsResearch ✓SMB ✓InformalEconomy✓Government✓ African Research Cen-tre on artificial intelli-genceTable 20: MI in Congo644.8.1 GovernmentOn February 24, 2022, the Economic Commission for Africa (ECA) and theGovernment of the Republic of Congo inaugurated a groundbreaking center de-voted exclusively to advancing research through MI, aiming to propel digitaltechnology in Africa across areas such as digital policy, infrastructure, finance,skills, digital platforms, and entrepreneurship. The African Research Centre onartificial intelligence, funded through the ECA and other partners, will providethe necessary technology education and skills to promote Africa’s integration,contributing to generating inclusive economic growth, stimulating job creation,breaking the digital divide, and eradicating poverty for the continent’s socio-economic development, ensuring Africa’s ownership of modern tools of digitalmanagement. The center was officially launched by the UN Under-Secretary-General and Executive Secretary of the ECA, and the Prime Minister of Congounder the auspices of President Denis Sassou Nguesso. The event was attendedby African ministers responsible for ICT and the digital economy. With the fullsupport of the Government of the Republic of Congo, the Center, the first ofits kind in Africa, will serve as a regional hub for the development of emergingtechnologies in the region. A partnership agreement to develop the project wassigned in March 2021 by the Republic of the Congo and ECA during the officialopening ceremony of the 7th session of the African Regional Forum on Sustain-able Development. UN partners include the United Nations Industrial Develop-ment Organization (UNIDO), UNESCO, the International TelecommunicationsUnion, Alibaba Jack Ma Foundation, and other key ECA partners. Congo willfunction as a regional MI hub across the continent, providing access to the deep-est and highest quality pool of MI talent. Aligned with Agenda 2063, the MICentre introduces a new dynamic to Africa’s participation in the global valuechain. Global companies choosing to locate in regional hubs can benefit fromstrong government support, low business costs, and access to world-class MIclusters. The MI Centre envisions working collaboratively with multiple stake-holders to establish linkages for a collaborative environment between industry,institutions, government, public and private sectors. The strategic pillars of thecenter include the provision of state-of-the-art MI research facilities, collabora-tion with top-ranked universities in Africa, building a network of highly skilledresearchers, and providing support and training to citizens to become scholars,researchers, leaders, and engaged individuals required to deliver digital transfor-mation in society. The African Research Centre on MI is now established in theDENIS SASSOU-N’GUESSO University of Kintélé, recognized as a platform forbusiness analysis on the continent. The MI Centre offers hybrid modes of train-ing in MI and robotics for researchers, youths, and interested citizens. It alsoprovides basic MI and Robotics skill-oriented training for talented elementaryand senior school students. The pursuit of a Master’s of Science Degree in MIand data science in collaboration with the University of Denis Sassou Nguesso isavailable at the MI Centre. With the aim of increasing the number of STEAM(Science, Technology, Engineering, and Mathematics) graduates and joining theranks of the world’s highest educated workforce, the depth and quality of the65MI Centre’s program content are remarkable. As part of the growing globaldigital and knowledge economy, the ECA works towards enabling countries topromote, adopt new and emerging technologies, and promote digital skills todeliver the transformation of their economies.4.9 Sao Tomé and Principe4.9.1 ResearchSao ToméandPrincipe2020-2023 Concrete ActionsResearch ✓ pepper value chainSMB ✓InformalEconomy✓Government✓Table 21: MI in Sao Tomé and Principe[121] develops Guiding mechanisms for Social Protection Targeting ThroughSatellite Data in Sao Tomé and Principe. Social safety net programs often tar-get the poorest and most vulnerable populations. However, in many developingcountries, there is a lack of administrative data on the relative wealth of thepopulation to support the selection process for potential beneficiaries of theseprograms. Therefore, the selection process often involves a multi-methodologicalapproach, starting with geographical targeting for the selection of program im-plementation areas.To facilitate this stage of the targeting process in Sao Tomé and Principe,their article develops High-Resolution Satellite Imagery (HRSI) poverty maps,providing estimates of poverty incidence and program eligibility at a highlydetailed resolution (110 m x 110 m). The analysis combines poverty incidenceand population density to facilitate the geographical targeting process.Their work demonstrates that HRSI poverty maps can serve as key oper-ational tools to aid decision-making in geographical targeting and efficientlyidentify entry points for rapidly expanding social safety net programs. UnlikeHRSI poverty maps based on census data, poverty maps based on satellite dataand machine learning can be updated frequently at a low cost, supporting moreadaptive social protection programs. [122] examines the role of value chainsanalysis in the agricultural sector, focusing on the case of pepper value chainsin Sao Tomé e Principe. Despite significant progress in yield and productivecapacity, global inequality persists, with hundreds of millions suffering fromhunger and undernourishment. Climate change poses a significant threat toagriculture, emphasizing the need for sustainable agricultural value chains thatpromote and protect natural resources while being inclusive and supportive of66smallholders. Value chain analysis, often used in policy development by NGOs,international organizations, and governments, plays a crucial role in promotingeconomic progress for less favored smallholders in the global market. In SaoTomé and Principe, where conditions such as dimension, insularity, and povertyare specific, the promotion of value chains is particularly dependent on foreignaid. The case of the pepper value chain reveals similarities between formal andinformal producers in terms of production systems, but the formal sector tendsto be more stable and financially protected than the informal sector.5 Western AfricaWestern Africa has a total area of 5,112,903 squared kilometers, with a pop-ulation of 418,544,337. The principal activities include agriculture, live stockmanagement, fishing, trade. In Western Africa, MI activities are diverse, withnotable research initiatives in Benin focusing on soil fertility assessment, ba-nana plant disease detection, electricity generation forecasting, public healthdecision-making for breast cancer, and suitability mapping for rice production.Burkina Faso is engaged in predicting malaria epidemics, mapping urban devel-opment, forecasting energy consumption, exploring mineral resources throughairborne geophysical data, and modeling monthly energy consumption. CaboVerde is involved in understanding aerosol microphysical properties, studyingclimate change’s impact on endemic trees, estimating salt consumption usingmachine learning, and monitoring volcanic eruptions. Côte d’Ivoire exploresmachine learning for cocoa farmers and progress towards onchocerciasis elim-ination. Gambia investigates machine learning models for pneumonia-relatedchild mortality and smart rural water distribution systems. Ghana’s MI re-search spans urban growth assessment, vehicle ownership modeling, public sen-timent analysis, blood demand forecasting, internet data usage analysis, sever-ity prediction of motorcycle crashes, effects of artisanal mining, and customsrevenue modeling. Guinea focuses on predicting viral load suppression amongHIV patients and prognosis models for Ebola patients. Guinea-Bissau delvesinto biomass relationships, cashew orchard mapping, learning and innovation insmallholder agriculture, and automatic speaker recognition for monitoring PL-HIV. Liberia engages in cloud computing and machine learning for land covermapping, predicting local violence, remote sensing for land cover studies, scal-able approaches for rural school detection, and open challenges for mapping ur-ban development. Mali’s research covers groundwater potential mapping, bore-hole yield predictions, cropland abandonment analysis, image recognition withdeep convolutional neural networks, MI’s role in addressing global health chal-lenges, improved recurrent neural networks for pathogen recognition, and mar-ket liberalization policy analysis. Mauritania explores MI-driven insights intoEnglish studies, desert locust breeding area identification, business intelligencemodels for e-Government, and remote monitoring of water points. Niger’s re-search includes electrical charge modeling, land use mapping using satellite timeseries, and adult literacy and cooperative training program analysis. Nigeria67showcases an extensive array of MI applications, including diabetes prevalencedetection, crude oil production modeling, flood area prediction, food insecurityprediction, entrepreneurial success prediction, mobile forensics for cybercrimedetection, genre analysis of Nigerian music, terrorism activity prediction, stockmarket forecasting, and poverty prediction using satellite imagery. Senegal’sresearch spans crop yield prediction, resilient agriculture, machine learning forrice detection, monitoring artisanal fisheries, predicting road accident sever-ity, estimating electrification rates, and analyzing the energy-climate-economy-population nexus. Sierra Leone collaborates with UNICEF’s Giga Initiative forrapid school mapping using MI and satellite imagery. Togo contributes to theNovissi program expansion, machine ethics, wind potential evaluation, maizeprice prediction, solar energy harvesting assessment, land use dynamics fore-casting, and solar energy harvesting evaluation. The government of Togo hoststhe Artificial Intelligence Week in 2024, emphasizing MI’s role in the country’sdevelopment. The humanitarian sector in Togo leverages machine learning algo-rithms and mobile phone data for effective aid distribution. Despite numerousstartups in Western Africa, [34] found that most of these countries are missingfrom the MI ecosystem. [35] reports the following number of companies thatspecialize in MI in Western Africa: Nigeria: 456, Ghana: 115, Ivory Coast: 29,Senegal: 23. [36] reports on the opportunities in agriculture. [37] investigatethe role of MI in SDGs from an African perspective.5.1 BeninBenin 2020-2023Concrete ActionsResearch ✓ soil, banana, rice,electricity, publichealthGovernment✓ SNIAM, SENIA, na-tional MI strategyTable 22: MI in Benin5.1.1 ResearchThe work in [462] assesses soil fertility status in Benin using digital soil mappingand machine learning techniques. Published in Geoderma Regional in 2022. Thework in [463] detects banana plants and their major diseases through aerial im-ages and machine learning methods, focusing on a case study in DR Congo andthe Republic of Benin. The work in [464] involves short-term electricity gener-ation forecasting using machine learning algorithms, with a case study of theBenin Electricity Community (CEB). The work in [465] utilizes mathematicalmodeling and machine learning for public health decision-making, with a focus68on the case of breast cancer in Benin. The work in [466] maps suitability for riceproduction in inland valley landscapes in Benin and Togo using environmentalniche modeling.5.1.2 GovernmentBenin has an institutional framework composed of several entities, includingthe Ministry of Digitalization, the Agency for Information Systems and Digital,the Authority for Personal Data Protection (APDP), and the Sèmè City Devel-opment Agency (ADSC). In addition, some private actors, primarily startupsand citizen associations, are already exploring avenues and organizing trainingon application development and MI usage. Benin also has training institutessuch as the Institute of Training and Research in Computer Science and the In-stitute of Mathematics and Physical Sciences equipped with a supercomputer,providing significant computing power for MI development in Benin and the en-tire sub-region. The Benin Government, through its action program, has madedigitalization a cornerstone of economic and social progress. Significant invest-ments in this sector since 2016 demonstrate a strong political will to develop thedigital economy and transform the country into a regional platform for digitalservices sustainably. Major digital sector projects (data centers, interoperabil-ity, e-Administration, e-Services, Open Data, etc.) will generate massive data,requiring proper management and utilization to ensure the creation of value re-mains within the Beninese economy. Machine intelligence emerges as a tool toeffectively address this issue and support Benin’s influence in strategic sectorssuch as education, health, agriculture, environment, and tourism. However, toharness its potential and strengthen existing initiatives in artificial intelligence,Benin needs to identify its strengths and define objectives to take a leader-ship position in the sub-region. To achieve this, the Ministry of Digitalizationhas initiated the development of the National Strategy for Machine Intelligenceand Big Data and its action plan, collaboratively designed with stakeholders andsubmitted for approval by the Council of Ministers. In its session on January 18,2023, the Council of Ministers approved the National Artificial Intelligence andBig Data Strategy (SNIAM) 2023-2027. Driven by the vision of making Benina country that shines by leveraging its massive data through MI systems andtechnologies, it consists of four programs implemented in three phases over fiveyears, with a portfolio containing one hundred twenty-three actions impactingthe public and private sectors. Its adoption positions Benin as a country capa-ble of seizing current and future opportunities related to artificial intelligenceand massive data processing, making it more attractive for investments fromthe private sector and development partners. With a projected amount of fourbillion six hundred eighty million (4,680,000,000) CFA francs over a five-yearperiod, the implementation of this strategy provides an opportunity to leverageMI in targeted development areas to position the country as a major player inMI in West Africa. For the second consecutive year, the Digital Entrepreneur-ship and Machine Intelligence Fair (SENIA) took place in Cotonou on May 12and 13, 2023. Bringing together nearly 1,000 participants, the event reflects69Benin’s ambition to become a major player in MI in West Africa. The debatesaim to generate impactful initiatives benefiting the government, private sector,and Beninese society as a whole. Organized by the Ministry of Digitalization,SENIA brings together Beninese talents and international experts specializingin artificial intelligence and data science. Through numerous conferences anddemonstrations, SENIA is not only a platform for exchange but also a businessand networking environment for industry professionals.With the goal of fostering research, education, and implementation of MIin Africa, an Artificial Intelligence Research Centre has launched in Cotonou,Benin Republic. Founded in March 2018, Atlantic AI Labs is attempting tounleash the potential of MI for sustainable development in healthcare, precisionagriculture, education, unmanned aerial vehicles, clean energy, environmentalprotection, and wildlife conservation.5.2 Burkina FasoBurkinaFaso2020-2023Concrete ActionsResearch ✓ CITADEL, urbansystems, energy,miningSMB ✓ African FoodsNutrition,DataBusiness-AI, CyberLabsTech, Kumakan ,SOSEB, Saintypay,Kalabaash, Dunia ,Qotto, Toto RiiboInformalEconomyGovernment✓ InterdisciplinaryCenter of Excel-lence in ArtificialIntelligence forDevelopmentTable 23: MI in Burkina Faso5.2.1 ResearchThe work in [457] focuses on predicting malaria epidemics in Burkina Faso us-ing machine learning. The work in [458] maps patterns of urban development70in Ouagadougou, Burkina Faso, utilizing machine learning regression model-ing with bi-seasonal Landsat time series. The work in [459] explores machinelearning models to predict town-scale energy consumption in Burkina Faso. Thework in [460] employs machine learning techniques on airborne geophysical datafor mineral resources exploration in Burkina Faso. The work in [461] focuses onforecasting models for monthly energy consumption using machine learning inBurkina Faso.5.2.2 Small BusinessesQotto sells solar home systems to rural households in West Africa with a pay-as-you-go model. At Qotto, we transform the daily life of our customers byinstalling solar kits. To help them achieve their dreams, we offer electric au-tonomy with great flexibility and quality customer service. Toto Riibo is aBurkina Faso-based online food ordering and delivery service. Dunia Paymentis an Ouagadougou-based mobile wallet startup that lets its users send and re-ceive money, pay in stores with a simple QR Code. African Foods Nutrition isan agri-food processing unit with an industrial focus. Their solutions involvethe innovative production of standard nutritional flours based on three typesof cereal and two legumes, nutritional flours enriched with Moringa, nutritionalflours enriched with baobab powder, nutritional flours enriched with dried fruits,nutritional flours enriched with other vegetables, nutritional flours for breast-feeding women, nutritional flours for pregnant women and anemic individuals,baby compotes, and natural wellness products. DataBusiness-AI is a consultingcompany specializing in MI solutions. CyberLabs Tech develops MI-enabledrobots for agriculture. Kumakan Studio creates games that promote Africanculture, myths, legends, and folklore to entertain and showcase African trea-sures to the world. Additionally, they create games for social change to helpsolve local problems. SOSEB, also known as SOS Energie Burkina, is a com-mercial enterprise with a social and environmental mission that aims to protectthe environment and promote sustainable development in Burkina Faso. In thisregard, it distributes solar electrification kits, solar irrigation kits, and ecologicalcooking solutions (such as solar ovens, biomass cookers, ecological coal, etc.).Saintypay is assisting businesses and users in facilitating money transfers be-tween Africa and Dubai. Kalabaash operates in the digital field and consists ofa team of dynamic experts with various skills. Kalabaash aims to be a leaderin the field of big data and artificial intelligence in Burkina Faso.5.2.3 GovernmentAfter inaugurating significant electronic communication infrastructure in Bobo-Dioulasso, on September 4, the country hosted the 16th edition of the DigitalWeek from September 8 to 10, 2020, in Ouagadougou, under the theme: ”Arti-ficial Intelligence: Opportunities and Challenges.” The Interdisciplinary Centerof Excellence in Artificial Intelligence for Development (CITADEL) will hostresearchers from Burkina Faso seeking a conducive environment for conduct-71ing high-quality, globally competitive, interdisciplinary research relevant to theAfrican context. CITADEL will also train new talents with versatile skills tomeet the needs of local industry and research. It is located at the Virtual Uni-versity of Burkina Faso and is endowed with funding of one million Canadiandollars.5.3 Cabo VerdeCaboVerde2020-2023Concrete ActionsResearch ✓ salt, climatechange, volcaniceruptions monitor-ingTable 24: MI in Cabo Verde5.3.1 ResearchThe work in [453] aims to understand aerosol microphysical properties basedon 10 years of data collected at Cabo Verde, utilizing an unsupervised machinelearning classification. The work in [454] explores the implications of climatechange on the distribution and conservation of Cabo Verde endemic trees. Thework in [455] involves the development, validation, and application of a machinelearning model to estimate salt consumption in 54 countries, including CaboVerde.The work in [456] utilizes Sentinel-1 GRD SAR data for volcanic eruptionsmonitoring, focusing on the case-study of Fogo Volcano in Cabo Verde during2014/2015.5.4 Côte d’IvoireIvoryCoast2020-2023 Concrete ActionsResearch ✓ CacaoSMB ✓ Futurafric AIGovernment✓ AI and Robotics Cen-ter in YamoussoukroTable 25: MI in Ivory Coast725.4.1 ResearchThe work in [451] investigates Machine Learning as a Strategic Tool for HelpingCocoa Farmers in Côte D’Ivoire. The work in [452] examines progress towardsonchocerciasis elimination in Côte d’Ivoire.5.4.2 GovernmentIn 2018, MainOne launched Abidjan data center, which offers capacity for 100racks. MainOne has launched a second Cote d’Ivoire data center in 2023. Thisnew Tier III-quality data center is in the Village of ICT & Biotechnology ofCote d’Ivoire (VITIB) in Grand Bassam, on the outskirts of Abidjan.In April 2023, the Ministry of Communication and Digital Economy of Côted’Ivoire, in collaboration with Smart Africa, inaugurated the Cybersecurity In-novation Center. Located at the African Higher School of Information andCommunication Technologies (ESATIC) in Abidjan, this center focuses on com-bating cybercrime. Aligned with the agreement signed in September 2022 withthe Republic of Côte d’Ivoire, the center aims to enhance digital skills, includingthe implementation of the Smart Africa Digital Academy (SADA). Supportedby ESATIC and Hitachi Systems Security Inc., this innovation center serves asa tool to improve national cybersecurity culture through awareness and skilldevelopment for the target population.There are several ongoing initiatives in Côte d’Ivoire to promote the de-velopment of MI. These include the Digital Transformation Initiative of Côted’Ivoire, creating an environment for MI development, and the Africa MI Ini-tiative, a partnership between the Ivorian government and the World Bank tofoster MI development. The country hosts universities like the University ofAbidjan and the MI and Robotics Center in Yamoussoukro actively involved inMI research and application development, particularly in areas like health, fi-nance, and agriculture. In 2022, a Franco-Ivorian collaboration was establishedthrough the Franco-Ivorian Hub for Education, launching the Master of ScienceBIHAR to support joint diploma programs between Ivorian and French institu-tions, emphasizing the crucial role of data in Africa’s digital future. The ESTIAaims to globally disseminate its Master of Science BIHAR through Digital As-sociate Campuses in partner universities for local tutoring of remote learners.In August 2023, the issue of MI in the context of Customer Experiencewas discussed during the second edition of LONACI Online mornings at theIvory Coast National Lottery (LONACI), at the lagoon hall of Ivoire TradeCenter in Abidjan-Cocody. LONACI Online mornings serve as a platform fordiscussions on digital topics by Lonaci, aiming to ”better connect” with itsclients. During this event, various speakers from Yadec Consulting, FuturafricArtificial Intelligence, and Willis Towers Watson shared their insights on thetopic of MI.In September 2023, accompanied by the UNESCO Assistant Director for So-cial and Human Sciences, the Minister of Good Governance and Anti-Corruptionchaired, in Abidjan-Plateau, the launch of the implementation of the recommen-73dation on the ethics of Machine Intelligence in Côte d’Ivoire. The event wasattended by representatives from various organizations, including the High Au-thority for Audiovisual Communication, the Commission for Access to PublicInformation and Public Documents, the Virtual University, the Association ofBloggers of Côte d’Ivoire, the National Union of Journalists of Côte d’Ivoire,and civil society. Given the scope of MI use and the main theme focusing ondigital issues, the Minister of Good Governance mentioned having enlisted thesupport of the Minister of Digital Economy to serve as Co-lead. In the imple-mentation of this recommendation, the ministerial department will focus on thegovernance and ethics of MI, while the Ministry of Communication and DigitalEconomy will address the technical aspects of MI use in various sectors.5.5 GambiaGambia 2020-2023 Concrete ActionsResearch ✓ waterSMB ✓InformalEconomy✓Government✓Table 26: MI in Gambia5.5.1 ResearchThe work in [449] examines how to deploy Machine Learning Models Using Pro-gressive Web Applications: Implementation Using a Neural Network PredictionModel for Pneumonia Related Child Mortality in The Gambia. The work in[450] investigates smart rural water distribution systems in the Gambia.5.6 Ghana5.6.1 ResearchThe work in [428] assesses urban growth in Ghana using machine learning andintensity analysis, with a focus on the New Juaben Municipality.The work in [429] models vehicle ownership in the Greater Tamale Area,Ghana, utilizing machine learning techniques.The work in [430] employs statistical analysis and machine learning to studypublic sentiment on the Ghanaian government.The work in [431] utilizes machine learning algorithms for forecasting andbackcasting blood demand data at Tema General Hospital in Ghana.The work in [432] conducts a historical analysis and time series forecastingof internet data usage and revenues in Ghana using a machine learning-basedFacebook Prophet model.74Ghana 2020-2023 Concrete ActionsResearch ✓ gold miningIndustry ✓ MiningSMB ✓ Diagnosify,Xpendly, Kwanso,DatawareTech,Khalmax Robotics,mNotify, Green-Matics, DigiExt,CYST, CRI, Hug-gle.care,QualiTraceInformalEconomy✓ energyGovernment✓ National Artificial In-telligence Center, Re-sponsible Artificial In-telligence LabTable 27: MI in GhanaThe work in [433] focuses on severity prediction of motorcycle crashes inGhana using machine learning methods. Published in the International Journalof Crashworthiness in 2020.The work in [434] explores the local effects of artisanal mining in Ghana,providing empirical evidence.The work in [435] introduces GC3558, an open-source annotated dataset ofGhana currency images for classification modeling.The work in [436] applies Ito calculus and machine learning for the projectionof forward US dollar-Ghana cedi rates.The work in [437] uses machine learning and Google Earth Engine to quan-tify the spatial distribution of artisanal goldmining in Ghana, focusing on theconversion of vegetation to gold mines.The work in [438] models customs revenue in Ghana using novel time seriesmethods.The work in [439] applies machine learning to analyze jump dynamics in USDollar-Ghana Cedi exchange returns.5.6.2 IndustryNeural networks have been used in the mining industry in [440, 441, 443, 444,445].755.6.3 Small BusinessesDatawareTech is a data analytics company with a mission to empower organi-zations to gain insights from data for strategic decision-making.Khalmax Robotics is an EdTech company that provides robotics and MIproducts in education.mNotify is an MI-powered Customer Engagement tool for SMEs, facilitatingexponential growth.GreenMatics develops affordable autonomous solar-powered agricultural robots.DigiExt provides technical services to rice farmers for optimal crop growth,resulting in cost savings.CYST is a software innovation company founded in 2013 and based in Ghana.It specializes in artificial intelligence to create simple and easy-to-use technologysolutions relevant to local markets while adhering to international standards.CYST also has a research arm called CRI (CYST Research Institute).Huggle.care uses Machine Intelligence to enhance how people find the bestcare for the symptoms they experience.The QualiTrace concept is built on the idea of traceability, allowing con-sumers to trace food produce back to the farm gates. QualiTrace is an Agri-Tech startup that utilizes track and trace technology to authenticate agriculturalinputs (such as seeds and fertilizers) and outputs. QualiTrace not only authen-ticates but also provides a clear, simple means by which players in any givensupply chain can trace products along the chain to the final consumer.Xpendly is an MI-powered startup that uses artificial intelligence to diag-nose skin diseases, predict the name of the disease, determine its severity level,and assign patients to pharmaceutical services or dermatologists. Xpendly isa personal finance management app that enables young African Millennials tomanage their finances in one place, build alternative credit profiles with theirfinancial activity, and access tailored financial products that help them save andinvest. A solution to prevent, track, and monitor road traffic accidents by offer-ing passengers, drivers, and regulators an app for identifying accident locationsand monitoring speed.5.6.4 Informal EconomyInformal enterprises learn how to produce goods and services through cumu-lative and diverse ways. However, there is limited empirical evidence on howlearning processes influence the innovation of informal enterprises in Africa.The paper [446] examines the effects of two learning processes (apprenticeshipand formal interactions) on the product innovativeness of informal enterprises inGhana. Employing unique survey data on 513 enterprises and the Type II Tobitmodel, our analyses revealed that apprenticeship, on the one hand, enhances thetechnological capability of enterprises leading to product innovativeness, whilecompetitive formal interactions, on the other hand, provide important marketfeedback that enhances the innovativeness of enterprises. In addition, financiallyconstrained informal enterprises that compete with formal enterprises in prod-76uct markets performed poorly with their new products, compared with theircounterparts who were not financially constrained. The work in [447] focuses oninformal energy consumption in Ghana. The work in [448] studies the driversof undeclared works using machine learning.5.6.5 GovernmentIn 2019, the government launched the National Artificial Intelligence Center topromote the development and adoption of MI in the country. The center istasked with developing policies, strategies, and frameworks that will guide thedevelopment and deployment of MI in Ghana.In 2022, the Kwame Nkrumah University of Science and Technology has beenawarded a grant to fund the establishment of a Responsible Artificial IntelligenceLab under the AI4D Africa Multidisciplinary Labs project initiated by Inter-national Development Research Centre. The Responsible Artificial IntelligenceLab is hosted at the Kwame Nkrumah University of Science and Technology inGhana. RAIL seeks to be a first step in establishing a sustainable approach tonurturing local talent to engage in multidisciplinary, responsible MI for devel-opment research and innovation with a focus on women and that that respondsto capacity requirements of the public and private sector.The Responsible AI Network - Africa was founded through a partnership be-tween the Faculty of Electrical and Computer Engineering at Kwame NkrumahUniversity of Science and Technology in Ghana and the Institute for Ethics inArtificial Intelligence at the Technical University of Munich in Germany: Theaim is to build a network of scholars working on the responsible developmentand use of AI in Africa.5.7 GuineaGuinea 2020-2023 Concrete ActionsResearch ✓ waterSMB ✓InformalEconomy✓Government✓Table 28: MI in Guinea5.7.1 ResearchThe work in [425] focuses on the development of machine learning algorithmsto predict viral load suppression among HIV patients in Conakry, Guinea.77The work in [426] transforms clinical data into actionable prognosis models,utilizing a machine-learning framework and a field-deployable app to predict theoutcome of Ebola patients.The work in [427] uses machine learning in epidemiology. It characterizesof risk factors related to the occurrence of pulmonary and extra pulmonarytuberculosis in the province of Settat.5.8 Guinea-BissauGuinea-Bissau2020-2023 Concrete ActionsResearch ✓ cashew, speaker recog-nitionSMB ✓InformalEconomy✓Government✓Table 29: MI in Guinea-Bissau5.8.1 ResearchThe work in [421] explores the relationship between above ground biomass andALOS PALSAR data in the forests of Guinea-Bissau. The work in [422] fo-cuses on mapping cashew orchards in Cantanhez National Park, Guinea-Bissau.The work in [423] delves into endogenous learning and innovation in Africansmallholder agriculture, drawing lessons from Guinea-Bissau. The work in [424]introduces an Automatic Speaker Recognition application for monitoring PL-HIV in the cross-border area between the Gambia, Guinea-Bissau, and Senegal.5.9 LiberiaLiberia 2020-2023 Concrete ActionsResearch ✓ anti-violence, landcover changeSMB ✓InformalEconomy✓Government✓Table 30: MI in Liberia785.9.1 ResearchThe work in [416] employs cloud computing and machine learning to supportcountry-level land cover and ecosystem extent mapping in Liberia and Gabon.The work in [417] investigates predicting local violence in Liberia using evi-dence from a panel survey.The work in [418] focuses on remote sensing, machine learning, and changedetection applications for land cover studies in Liberia.The work in [419] presents a scalable approach using convolutional neuralnetworks and satellite imagery for detecting rural schools in Africa.The work in [420] introduces an open machine learning challenge to mapurban development and resilience in diverse African cities from aerial imagery.5.10 MaliMali 2020-2023 Concrete ActionsResearch ✓ Predominantly OralLanguages, MI inAfrica in 20 Questions,image recognition,farmers insurance,SMB ✓ Guinaga, Grabal,TimadieArt &MusicSK1 ART, Dronegra-phyInformalEconomy✓ WETEGovernment✓ CIAR-MaliTable 31: MI in Mali5.10.1 ResearchA recent book [3] titled ”MI in Africa in 20 Questions” with 11 co-authors fromAfrica addresses key questions raised by MI in Africa was published in June2023:• Question 1: The first question addressed is that of the definition of in-telligence, natural intelligence, human intelligence, machine intelligence,artificial intelligence, and collective intelligence.• Question 2: In an African context, the issue of generative machine intelli-gence emerges with particular importance, especially on online platforms.The authors of this book prompt us to reflect on the opportunities andchallenges this approach may present for Africa, addressing issues related79to the utopia and opportunity of generative machine intelligence in varioussectors such as the economy, finance, and agriculture.• Question 3: Conversational agents generate keen interest. The book ex-amines potential errors of these conversational agents and the need tocorrect them to ensure better interaction with African users.• Question 4: The book also tackles intriguing questions about delusions,illusions, confabulations, and hallucinations in generative machine intelli-gence. Do these phenomena actually exist or are they just analogies?• Question 5: Machine intelligence also finds applications in the financialsector in Africa. The book explores opportunities and challenges relatedto using machine intelligence for banking without traditional banks in theinformal sector.• Question 6: In the agricultural sector, can machine intelligence be consid-ered a delusion or an opportunity for Africa? This book guides us througha thorough reflection on the use of machine intelligence in African agri-culture, examining potential advantages such as cost reduction and opti-mization of agricultural practices.• Question 7: Manure is becoming increasingly interesting for mass agricul-ture in Africa. This book invites you to explore this alternative with ma-chine intelligence from plants and animal waste, particularly cattle urine.• Question 8: In the livestock sector, can machine intelligence be an ally inherd monitoring and improving animal welfare? This book encourages usto ponder the effects of machine intelligence in livestock.• Question 9: A crucial question arises: can the use of machine intelligencecontribute to reducing the cost of animal feed in Africa? The authorsof this book explore the possibilities offered by machine intelligence tooptimize animal feeding processes, highlighting the economic and envi-ronmental implications of this approach. What is the optimal strategybetween producing livestock feed oneself, arranging with local producers,and buying imported concentrates from other continents?• Question 10: The book also raises questions about access to machineintelligence in African contexts where internet access may be limited. Howcan the integration of machine intelligence be envisioned in environmentswhere connectivity is a major challenge?• Question 11: Another essential question addressed in this book concernsempowering women through machine intelligence. This book urges us toreflect on the socio-economic implications of this technology, evaluatingwhether its adoption can strengthen or, conversely, amplify existing in-equalities.80• Question 12: The book also examines the risks of inequalities and stereo-types arising from the massive use of machine intelligence. How can weensure that machine intelligence systems do not reproduce biases and dis-criminations present in African society? This book pushes us to questionpractices and responsibilities related to machine intelligence in the fightagainst inequalities and stereotypes.• Question 13: The spread of misinformation is a major challenge in thedigital world. The book explores the possibility that machine intelligencemay accentuate this phenomenon and highlights measures to mitigate thisrisk in Africa.• Question 14: In the fight against crime, can machine intelligence be aneffective tool, or does it represent a double-edged sword? The book invitesus to reflect on the implications for security, confidentiality, and humanrights in using machine intelligence to combat crime.• Question 15: How can machine intelligence contribute to civil protectionin Africa? The book examines the potential applications of this technologyin disaster prevention, crisis management, and improving the resilience ofAfrican communities.• Question 16: The health sector is also explored in the book. How canmachine intelligence be used to improve healthcare in Africa? This bookraises crucial questions about potential illegal practices and ethical issuesrelated to the use of machine intelligence in the healthcare field.• Question 17: Can the massive integration of machine intelligence poseenergy supply problems in certain African countries? This book leads usto reflect on the implications of this technology on electricity demand andto identify sustainable solutions to address this challenge.• Question 18: In the field of road safety, can the widespread use of machineintelligence lead to greater inattention, distraction, and insecurity? Thebook raises important questions about the potential consequences of inte-grating machine intelligence into vehicles and the need to ensure increasedsafety on African roads.• Question 19: The nineteenth question in this book discusses the potentialoffered by intelligence for auditing, but its reliability depends on severalfactors, including the quality of data, the accuracy of algorithms, and theability to adapt to changes.• Question 20: The integration of machine intelligence into education sparksmany discussions about the role of machines in teaching and learning. Thistwentieth and final question, far from the least, explores the question ofwho should be responsible for the use of machine intelligence in educationand what the proposed content will be. The use of machine intelligencein education can have several advantages, such as personalized learning,81providing instant feedback to students, and access to online learning re-sources. However, it is important to consider who will be responsible fordesigning and implementing these technologies.Most languages of the world are predominantly oral, with little to no writingtradition. In recent years, the African continent was completely missing on theNLP map, but due to efforts of grassroots communities such as Masakhane,Africa is now present on the NLP map focused on African languages. Despiteall the efforts made thus far, more progress is needed. As of August 2023, Mali’spopulation stands at approximately 22 million individuals. The country boasts13 national languages and an illiteracy rate of 65%. Interestingly, 80% of thepopulation communicates in Bambara. In the past, French held the position ofthe official language; however, merely 35% of the people were adept in functionalliteracy in French. Notably, Mali has recently embraced a new constitution thatrelegates French from its former status while elevating all 13 national languagesto the rank of official languages.DONIYA-SO a legal non-governmental platform on Data Science and MIin Bamako, Mali is studying B.A.M.B.A.R.A: Breaking Audio MultilingualBarriers Advancing Research Across Africa. The MI project aims to buildhigh-quality audio/sound/speech datasets for Predominantly Oral Languages(POLs). The work seeks to narrow this language gap by gathering top-notchaudio/sound/speech datasets for Bambara. This effort will not only facilitateresearch outcomes for the remaining 12 languages of Mali but also extend itsbenefits beyond. The state-of-the-art machine learning including deep learningmodels requires a significant amount of data, which is not readily available inlow-resource settings, and this is the case for POLs. If funded, this project hasthe potential to increase the amount of data usable by researchers, and develop-ers in the field of ML. This, when successfully implemented, has the potentialto increase literacy rate, open our people to the world, and the world to ourpeople.The work in [404] focuses on preprocessing approaches in machine-learning-based groundwater potential mapping in Mali, specifically the Koulikoro andBamako regions. The work in [405] presents multiclass spatial predictions ofborehole yield in southern Mali using machine learning classifiers. The work in[406] assesses cropland abandonment from violent conflict in central Mali usingSENTINEL-2 and Google Earth Engine. The work in [407] introduces a deepconvolutional neural network for image recognition in Mali. The work in [408]explores the potential of artificial intelligence in addressing global health chal-lenges, focusing on antimicrobial resistance and the impact of climate change ondisease epidemiology. The work in [409] utilizes an improved recurrent neuralnetwork with LSTM for the recognition of pathogens through image classifica-tion. The work in [410] analyzes three decades of market liberalization policyin Mali, specifically focusing on grain markets. The work in [411] introducesSatDash, an interactive dashboard for assessing land damage in Nigeria andMali. The work in [412] reflects on the role of digital technology in peace pro-cesses, emphasizing the need to make peace with uncertainty. The work in [413]82evaluates machine learning and deep learning classifiers for offensive languagedetection in a code-mixed Bambara-French corpus. The work in [414] presentsa typology of Malian farmers and their credit repayment performance usingan unsupervised machine learning approach. The work in [415] applies princi-pal component analysis to hydrochemical data from groundwater resources inBamako, Mali.5.10.2 Small BusinessesThere are also opportunities for private sector investment in MI-based solutionsfor agriculture in Mali. Startups and tech companies are developing innovativesolutions to address challenges faced by farmers. For instance, Timadie is an in-novative platform of platforms that is based on graphchain technologies adaptedfor less connected environments.In the agriculture sector, the Guinaga platform has physically engaged withover 400 farmers and is developing a graphchain as a traceability tool for foodproducts. Graphchains such as MangueChain, Karitéton, Yuton, and Riztonhave contributed to the valorization of local products in Senegal, Mali, andBurkina Faso. It utilizes MI, Blockchain, and Graphchain technologies to opti-mize the supply chain, air pollution, and understand production, consumption,sales, purchases, transportation, storage, and more in Mali, Burkina Faso, andSenegal. The data contributes to building a Kariteton, incorporating shea but-ter production. MI techniques aid in estimating production 8 weeks in advanceto reduce demand-supply mismatch. Guinaga also employs deep learning in itsManguechain for tracking mangoes from tree to consumer tables. Guinaga’sknitting and crochet club is a dedicated platform for cotton and indigo plantcultivation for natural dyes, semi-automatic machine knitting, providing a spacefor knowledge sharing, creativity, and support for knitting enthusiasts.Since 2019, SK1 Sogoloton uses deep learning for videos analytics for infor-mation dissemination in 25 languages across Africa. Users can access reliableand verifiable information, participate in information production, and combatmisinformation in Africa. The platform has over 10 million video views and 200million interactions as of November 2023. CI4SI, from the learning and gametheory laboratory, emphasizes collective intelligence and encourages collabora-tion to solve societal problems.The Grabal platform is a catalyst for connecting traditional livestock breed-ers in Africa, promoting the preservation of genetic diversity. MoutonChain,in particular, has ensured the traceability of sheep, providing consumers withthe necessary confidence during the Tabaski period from 2019 to 2023. Graballinks breeders directly to buyers from the field, utilizing MI-enabled drones toestimate animal food, headcount, and safe paths in Mali, Burkina Faso, andSenegal.WETE has opened new perspectives for the economic empowerment ofwomen in Africa. To date, 150 African women CEOs are present on the platform.Women-in-Drones promotes the participation of women in emerging sectors suchas drone technology and dronegraphy. To date, 23 women have been trained in83drone piloting and the uses of MI in video-processing in Mali. The particularityhere is that these trained women uses MI-enabled drones in their professionalworks such as mineral extraction, geo-information, and crisis management.By fostering interconnectivity and encouraging co-opetition, Timadie pro-vides an environment conducive to the emergence of new ideas, fruitful collab-orations, and significant social innovations in multiple African countries.Timadie, hosting various platforms, has trained students at several schoolson MI applications in geo-information systems, economy, transportation, energy,healthcare, and agriculture.5.10.3 Art & MusicSK1’ART tests the accuracy and reliability of several music and Dogon maskgenerated by MI. Together with Malian artists the platform aims to be createan online Malian Got Talent combined with MI and blockchain technologies.5.10.4 GovernmentThe groundbreaking ceremony on June 7, 2023, marked a crucial moment forMali’s recent technological ambitions. Colonel Assimi Goita, leading the tran-sitional government, initiated the construction of the Center for Artificial In-telligence and Robotics (CIAR-Mali) in Kati, Koulikoro region in Mali. Thisvisionary project, designated as a Public Scientific, Technical, and Cultural Es-tablishment, is set to become a hub for cutting-edge research, development, andeducational activities in artificial intelligence and robotics. With an estimatedconstruction cost of 3.3 billion F CFA (4.5 million euros), the CIAR-Mali un-derscores the government’s firm belief in the transformative power of artificialintelligence. This strategic investment not only signifies a commitment to tech-nological advancement but also presents an unparalleled opportunity for theyouth of Mali and Africa. Aspiring individuals keen on contributing to the con-tinent’s technological evolution now have a potential avenue through the CIAR-Mali. Fast forward to October 19, 2023, the ratification bill for the creation ofthe CIAR-Mali received the unanimous approval of the National TransitionalCouncil. This legislative approval solidified the center’s role in fostering inno-vation and knowledge transfer within the domains of artificial intelligence androbotics. However, despite the legislative green light, the center has not yetcommenced its operational phase as of the current date. The CIAR-Mali un-derscores the nation’s forward-looking approach, emphasizing the pivotal rolethat artificial intelligence and robotics play in shaping the future of Mali andcontributing to the broader technological landscape of the African continent.5.11 Mauritania5.11.1 ResearchThe work in [400] explores MI-driven insights into the factors influencing stu-dents’ choice of English studies as a major at the University of Nouakchott84Al Aasriya, Mauritania. Published in the International Journal of Technology,Innovation, and Management in 2022.The work in [401] leverages MI methodologies for identifying desert locustbreeding areas in Mauritania through Earth Observation.The work in [402] employs MI techniques to survey business intelligence mod-els for e-Government in Mauritania. The authors, El Arby Chrif et al., showcasethe MI applications at The International Conference on Artificial Intelligenceand Smart Environment in 2022.The work in [403] utilizes MI and IoT (LoRa) technology for the remotemonitoring of water points in Mauritania. This AI-infused study is publishedin the Indonesian Journal of Electrical Engineering and Informatics in 2022.5.12 NigerNiger 2020-2023 Concrete ActionsResearch ✓ electricity, trainingSMB ✓InformalEconomy✓Government✓Table 32: MI in Niger5.12.1 Research[397] examines Electrical Charge of Niamey City using a neural network model.[398] studies land use mapping using Sentinel-1 and Sentinel-2 time series in aheterogeneous landscape in Niger, Sahel.[399] examines adult literacy and cooperative training programs in Niamey,Niger.5.13 Nigeria5.13.1 ResearchThe work in [366] by Oyebode and Orji focuses on detecting factors responsiblefor diabetes prevalence in Nigeria using social media and machine learning.Obite et al. contribute to the modeling of crude oil production in Nigeria,identifying an eminent model for application [367]. Ighile et al. apply GISand machine learning to predict flood areas in Nigeria [368]. Villacis et al.explore the role of recall periods in predicting food insecurity in Nigeria usingmachine learning [369]. McKenzie and Sansone analyze the competition betweenman and machine in predicting successful entrepreneurs in Nigeria [370]. Goniand Mohammad present a machine learning approach to a mobile forensics85Nigeria 2020-2023 Concrete ActionsResearch ✓ PatatoSMB ✓ Tuteria, Kudi AI , Cu-racel , Codar TechAfricaArt &Music✓ Afrobeats, InfiniteEchoesInformalEconomy✓Government✓ NCAIRTable 33: MI in Nigeriaframework for cybercrime detection in Nigeria [371]. Folorunso et al. dissect thegenre of Nigerian music with machine learning models [372]. Lawal et al. predictfloods in Kebbi state, Nigeria, using machine learning models [373]. Nwankwoet al. focus on predicting house prices in Lagos, Nigeria, using machine learningmodels [374]. Panjala et al. identify suitable watersheds across Nigeria usingbiophysical parameters and machine learning algorithms for agri-planning [375].McKenzie and Sansone, in [376], presents the challenges of predicting en-trepreneurial success, drawing evidence from a business plan competition inNigeria. Gladys and Olalekan present a machine learning model for predictingcolor trends in the textile fashion industry in southwest Nigeria [377]. Odeniyiet al. predict terrorist activities in Nigeria using machine learning models [378].Ogundunmade and Adepoju model liquefied petroleum gas prices in Nigeria us-ing time series machine learning models [379]. Ekubo and Esiefarienrhe utilizemachine learning to predict low academic performance at a Nigerian university[380]. Muhammad and Varol propose a symptom-based machine learning modelfor malaria diagnosis in Nigeria [381]. Salele et al. model run-off in perviousand impervious areas using SWAT and a novel machine learning model in CrossRiver State, Nigeria [382]. Adeeyo and Osinaike model the oil viscosity of Nige-rian crudes using machine learning [383]. Jean et al., in [384], combine satelliteimagery and machine learning to predict poverty in Nigeria. Ibrahim et al. pre-dict potato diseases in smallholder agricultural areas of Nigeria using machinelearning and remote sensing-based climate data [385]. Oyebode and Orji, in[386], explore social media and sentiment analysis in the context of the Nigeriapresidential election 2019. Eneanya et al. examine environmental suitabilityfor lymphatic filariasis in Nigeria [387]. Obulezi et al. predict transportationcosts inflated by fuel subsidy removal policy in Nigeria using machine learning[388]. Mbaoma et al. use geospatial and machine learning-driven air pollutionmodeling in Agbarho, Delta State, Nigeria [389]. Achara et al. investigate fi-nancial institution readiness and adoption of machine learning algorithms andperformance of select banks in Rivers State, Nigeria [390]. Oyewola et al. pro-pose a new auditory algorithm for stock market prediction in the Nigerian stock86exchange, focusing on the oil and gas sector [391].The work in [387] by Oyebode and Orji explores social media and senti-ment analysis, specifically examining the Nigeria presidential election in 2019.Eneanya et al. in [393] investigate the environmental suitability for lymphaticfilariasis in Nigeria. Obulezi et al. [394] focus on machine learning models forpredicting transportation costs inflated by fuel subsidy removal policy in Nige-ria. Mbaoma et al. [395] utilize geospatial and machine learning-driven airpollution modeling in Agbarho, Delta State, Nigeria. Achara et al. [396] studyfinancial institution readiness and the adoption of machine learning algorithms,examining the performance of select banks in Rivers State, Nigeria. Oyewola etal. [392] introduce a new auditory algorithm for stock market prediction in theNigerian stock exchange, particularly emphasizing the oil and gas sector5.13.2 Small BusinessesTuteria is an online platform that connects people who are seeking to learn any-thing with those who live near them and are available to teach them. It providesan environment that offers safety, accountability and quality. Globally, conven-tional methods of education and learning are being challenged. They are movingfrom centralized to distributed, from standardized to personalized. These trendshave repeatedly demonstrated their ability to deliver better learning outcomesand Tuteria fits in well with this trend.Kudi AI has a familiar origin as Kudi means money in the Hausa language,reflecting its simplicity in service options. Kudi is a conversational agent pow-ered by machine intelligence, based in Nigeria, that assists with your financessimply by asking. Kudi uses a conversational machine intelligence system tointeract with you on a daily basis. It helps you transfer money, track your ac-count details, purchase airtime, pay recurring bills, and also reminds you whensome of these bills are due.Codar Tech Africa in Lagos, Nigeria was created in December 2021 realizingthat tech training was inadequate in Africa’s biggest economy. With the swiftly-changing job market because of advancements in generative AI, there is animportance and urgency of teaching tech skills that would set people apart inan AI-influenced landscape. To date, Codar Africa provides hands-on trainingexperience in various aspects of tech including; data analysis, web developmentand design, Search Engine Optimization, and Cybersecurity.Nigerian-based Curacel is an MI platform that aims to drive insurance pen-etration in emerging markets via APIs enabling insurers to connect with dig-ital distribution channels and administer their claims. Founded in 2019, theMI platform recently raised $3 million in seed funding. Initially the platformwas intended to be an electronic health information management platform forhealthcare providers, enabling clinics to digitize and manage paper records, ap-pointments, patient communications, billing and reporting through a web app.Ubenwa Health is a MedTech startup in Nigeria building the future of auto-mated sound-based medical diagnostics.875.13.3 Art & MusicThe Afrobeats music genre from Lagos has recently kept millions on their feetand challenged preconceptions of African music. Its popularity is growing glob-ally, with an increasing number of players in the entertainment industry tryingto get a slice of it. When machine intelligence apps began spreading in Nigeria’smusic industry, Eclipse Nkasi thought his days as a producer were numbered.However, he took a step back, identified opportunities as well as threats, andused the technology to generate a whole new Afrobeats album in his studio onthe outskirts of Lagos. MI doesn’t have to replace what we have. It gives peoplea new experience... and that’s how I believe MI is really going to shake things.In the past, it would have taken him thousands of dollars and up to three monthsto compose the tracks, recruit musicians, record performances, refine them ina traditional studio, and release them to fans. Nigerian artists activated MIalgorithms and set them to work, assisting in creating the nine-track album”Infinite Echoes.” They instructed it to auto-generate song lyrics and titles,including ”God Whispers,” ”Love Tempo,” and ”Dream Chaser.” Then, theymodified the words themselves to fit their chosen theme - a struggling artistnot giving up on their passion for creating music. Next, they used another MItool to generate the tunes. Nkasi recorded some vocals and fed them into yetanother app, transforming his vocals into the voice of the album’s generatedsinger - a virtual ”singer” named Mya Blue, who appears online as a computeranimation in front of her audience. Certain things may become obsolete dueto MI, but it should also create opportunities for artists to reinvent themselvesand improve their work more efficiently. The technology is already transformingthe industry and could have a positive impact on production values and othertechnical aspects of the recording process. However, there are still uncertaintiesand areas, including copyright, that need consideration and development.Initially launched in February 2023 in the US and Canada, followed by theUK and Ireland in May, DJ is a personalized MI guide that offers users a care-fully curated lineup of music, accompanied by commentary on the tracks andartists, all delivered in a realistic voice. The feature aims to foster a strongerconnection between users and their music, enabling discovery through tailoredrecommendations. The initial voice model for DJ was based on Spotify’s Headof Cultural Partnerships, Xavier ’X’ Jernigan. However, the latest rollout ex-tends the offering to include commentary in English for listeners in variousinternational markets. African countries where the DJ feature is now accessi-ble include Botswana, Burundi, eSwatini, the Gambia, Ghana, Kenya, Lesotho,Liberia, Malawi, Namibia, Nigeria, Rwanda, Sierra Leone, South Africa, Tan-zania, Uganda, Zambia, and Zimbabwe. Never before has listening felt so com-pletely personal to each and every user, thanks to the powerful combination ofSpotify’s personalization technology, generative AI, and a dynamic, expressivevoice. Users can access the AI DJ by opening the Spotify mobile app on theiriOS or Android devices. After launching the app, they can navigate to theMusic Feed on the homepage and tap play on the DJ feature.885.13.4 GovernmentThe National Centre for Artificial Intelligence and Robotics (NCAIR) is one ofNITDA’s special purpose vehicles created to promote research and developmenton emerging technologies and their practical application in areas of Nigeriannational interest. The center, a state-of-the-art facility, along with its moderndigital fabrication laboratory, is co-located in the same building complex withthe Office for Nigerian Digital Innovation, at No. 790 Cadastral Zone, WuyeDistrict, Abuja. NCAIR as a digital innovation and research facility is focusedon Artificial Intelligence, Robotics and Drones, Internet of Things, and otheremerging technologies, aimed at transforming the Nigerian digital economy, inline with the National Digital Economy Policy and Strategy. NCAIR is also fo-cused on creating a thriving ecosystem for innovation-driven entrepreneurship,job creation and national development. In 2023, Nigeria is extending an invi-tation to scientists of Nigerian heritage, as well as globally renowned expertswho have worked within the Nigerian market, to collaborate in the formula-tion of its National Artificial Intelligence Strategy. According to the Ministerof Communications, Innovation and Digital Economy, the National InformationTechnology Development Agency has initiated the development of a NationalMI Strategy. The action will impact the way the government formulates newtechnological solutions for its critical national challenges. As a result, the gov-ernment is broadening its co-creation strategy by assembling a selection of lead-ing MI researchers with Nigerian heritage from around the world. The Nigeriangovernment recognizes that MI has developed into a versatile technology, re-shaping production and services and holds immense potential for influencingsocietal progress and economic expansion. According to a white paper titledCo-creating a National Artificial Intelligence Strategy for Nigeria, a sophisti-cated method was used to pinpoint accomplished MI researchers with Nigerianroots, using global MI publication data and advanced machine learning models.Research index was created to locate influential machine intelligence researchersof Nigerian heritage. As the preliminary research phase concludes, the Nigeriangovernment seeks public involvement, acknowledging the potential for errorsand aiming to tap into collective knowledge and insights. The government ofNigeria via the National Centre for Artificial Intelligence and Robotics aims toestablish communities of MI developers nationwide to influence the country’stechnological future. The initiative started in three states in 2023, followed bystrategic planning for its extension to additional states and, eventually, all localgovernment areas.5.14 Senegal5.14.1 ResearchThe work in [346] by Sarr and Sultan predicts crop yields in Senegal using ma-chine learning methods, focusing on climatology. In [347], Nyasulu et al. con-tribute to resilient agriculture in the Sahel region, employing machine learningfor weather prediction. MMbengue et al. evaluate machine learning classifica-89Senegal 2020-2023 Concrete ActionsResearch ✓ mangoSMB ✓ Afrikamart , TerangaCapital,InformalEconomy✓ Lengo AIGovernment✓ Data Center, NationalMI strategyTable 34: MI in Senegaltion methods for rice detection using Earth observation data in Senegal [348].Bayet et al. apply a machine learning approach to enhance the monitoring ofSustainable Development Goals, concentrating on Senegalese artisanal fisheries[349]. Dia et al. present a hybrid model for predicting road accident severity inSenegal, utilizing artificial intelligence and empirical studies [350]. The work in[351] by State et al. explores explainability in practice by estimating electrifica-tion rates in Senegal using mobile phone data. In [352], Sarkodie et al. conductan empirical analysis of the energy-climate-economy-population nexus in Sene-gal and other countries. Kebe et al. share their experience with detecting andclassifying Quality-Of-Service problems in MV/LV distribution substations inSenegal using artificial intelligence [353]. Seck and Diakité develop supervisedmachine learning models for predicting renal failure in Senegal [354]. Lee etal. investigate the intersection of colonial legacy and environmental issues inSenegal through language use [355]. The work in [356] by Dione et al. focuseson designing Part-of-Speech-Tagging resources for Wolof in Senegal. Alla et al.leverage LSTM to translate French to Senegalese local languages, with Wolof asa case study [357]. Dia et al. present an empirical study on predicting the sever-ity of road accidents in Senegal using a hybrid model [358]. Dione et al. proposean IoT-based e-health model for developing countries, with Senegal as a casestudy [359]. Özdogan and Govind examine three decades of forest cover changein Senegal using remote sensing [360]. Traoré et al. analyze nonlinear pricetransmission in the rice market in Senegal, employing a model-based recursivepartitioning approach [361]. Drame et al. conduct an analysis and forecast ofenergy demand in Senegal using SARIMA and LSTM models [362]. Sarron etal. investigate the efficiency of machine learning for mango yield estimation inSenegal under heterogeneous field conditions [363]. Diop et al. apply a machinelearning approach to the classification of Okra [364]. Moustapha Mbaye et al.propose a new machine learning workflow for creating an optimal waiting list inhospitals [365].5.14.2 Small BusinessesAfrikamart is an agritech that facilitates the pick-up, shipping, and trading offresh fruits and vegetables between small producers and urban retailers via a90digital platform. Afrikamart was founded to address agricultural loss issuespresent throughout the production chain, from producer to retailer. Afrikamartis supported by Acceleration Technologies, a 2.5 million euros program that aimsto finance and support fifteen digital start-ups in Sub-Saharan Africa, supportedby AFD through the Digital Africa initiative. In Senegal, Teranga Capital is incharge of implementing this program.5.14.3 Informal EconomyLengo AI is the first AI-Driven Intelligence Platform for for Fast Moving Con-sumer Goods companies in Africa for the Informal Sector.5.14.4 GovernmentIn December 2016, during the State visit to France by President Macky Sall,Ministers of Economy, Finance, Planning, and Higher Education and Researchwere instructed to sign a 15 million euros financing deal with the Public In-vestment Bank (BPI). This funding aimed at establishing the National Cen-ter for Scientific Computing (CNCS) in Diamniadio, equipped with a paral-lel calculator, the most powerful south of the Sahara excluding South Africa.The funding covered mobility, maintenance, and training, benefiting researchin artificial intelligence, Big Data, cybersecurity, robotics, and scientific com-puting. The CNCS, operational today, contributes to sectors like agriculture,health, genomics, biotechnology, ICT, mining, gas, oil, energy, security, mete-orology, climate change, coastal erosion, water management, navigation, andenvironmental data exploitation. The supercomputer, with a computing powerof around 320 Tflops and integrated storage capacity of 21 Terabytes, alignswith the Senegal Emergent Plan in agrohydrology and the mining sector, facil-itating numerical simulations for innovation in meteorology, climatology, imageprocessing, vegetation growth, and mineral exploration.In May 2021 was the inauguration of the Data Center of Diamniadio, Sene-gal. This infrastructure is a major revolution for the digital sovereignty of ourcountry, which for the first time will have its own digital data storage struc-ture. Among other features, the Data Center will make it possible to generalizevery high speed across the national territory, to satisfy, at affordable costs,government and private sector requests for hosting and operation of computerplatforms and data. This Data Center will also facilitate the dematerializationof procedures.In May 2022, An experts consultative meeting on developing a continentalstrategy for Machine Intelligence in Africa was successfully held in Dakar, Sene-gal. This was held on the margins of the 6th Calestous Juma Executive Dialogue(CJED), organized by the African Union High-Level Panel on Emerging Tech-nologies (APET). The CJED convenes policy and decision makers, executives,youth, and relevant stakeholders to deliberate on harnessing appropriate innova-tions and emerging technologies for Africa’s socio-economic development. APEThas prioritized and recommended MI as an emerging technology worth harness-91ing for Africa’s socio-economic development. In the APET ”AI for Africa”report launched in December 2022, the panel provides guidelines for Africancountries on how best to exploit AI-based technologies for the continent’s ad-vancement. The high-level panel further recommended developing a continentalMachine Intelligence strategy for Africa, necessitating this expert consultativemeeting.In June 2023, Senegal developed its National MI Strategy. This strategy,developed through extensive stakeholder consultations, aimed to position Sene-gal as an MI leader in West Africa. A delegation composed of members fromthe Ministry of Communications, Telecommunications and Digital Economy ofSenegal, academia and the private sector embarked on a study tour to Rwanda,an African pioneer in digital innovation. This tour, a strategic move in op-erationalizing Senegal’s National MI strategy, was not just about technologytransfer but a deeper dive into the complexities of MI in an African context.The initiative, led by Enabel, one of the implementing partners of the AU-EUD4D Hub Project, aimed to gather insights and best practices to operationalizeSenegal’s National MI Strategy.5.15 Sierra LeoneSierraLeone2020-2023 Concrete ActionsResearch ✓SMB ✓InformalEconomy✓Government✓Table 35: MI in Sierra Leone5.15.1 ResearchUNICEF’s Giga Initiative endeavors to map every school on the Planet. Know-ing the location of schools is the first step to accelerate connectivity, onlinelearning, and initiatives for children and their communities, and drive economicstimulus, particularly in lower-income countries. Development Seed [345] isworking with the UNICEF Office of Innovation to enable rapid school mappingfrom space across Asia, Africa, and South America with AI. In seven monthsof development and implementation, we added 23,100 unmapped schools to themap in Kenya, Rwanda, Sierra Leone, Niger, Honduras, Ghana, Kazakhstan,and Uzbekistan. To accomplish this we built an end-to-end scalable MI modelpipeline that scans high-resolution satellite imagery from Maxar, applies ourhighly refined algorithm for identifying buildings that are likely to be schools,and flags those schools for human review by a Data Team.925.16 TogoTogo 2020-2023 Concrete ActionsResearch ✓ maize, MI ethicsSMB ✓ Semoa, Eazy Chain,SocialGIS, DobbeePay, Solimi Fintech ,ArtybeGovernment✓ GiveDirectly, AfricanCybersecurity Re-source Center, Artifi-cial Intelligence WeekTable 36: MI in Togo5.16.1 ResearchThe research efforts in [338] focused on helping the government expand theNovissi programme from informal workers in Greater Lomé to poorer individualsin rural regions of the country, and were designed to meet the government’stwo stated policy objectives: first, to direct benefits to the poorest geographicregions of the country; and second, to prioritize benefits to the poorest mobilesubscribers in those regions. Individuals without access to a mobile phone couldnot receive Novissi payments, which were delivered digitally using mobile money.The approach they developed, uses machine learning to analyze non-traditionaldata from satellites and mobile phone networks. The work in [339] by Kohnert(2022) explores machine ethics and African identities, offering perspectives onthe role of artificial intelligence in Africa. The work in [340] assessed windpotential in Togo’s Kara region using artificial neural networks, offering bothstatic and dynamic evaluations. The work in [341] focused on predicting maizeprices in Lome, Togo, utilizing a Hidden Markov Chain Model. The work in[342] compared survey-based impact estimation with digital traces in the contextof randomized cash transfers in Togo. The work in [343] forecasted land use andcover dynamics in the Agoènyivé Plateau, Togo, using a combination of remotesensing, machine learning, and local perceptions. The work in [344] employedartificial neural networks to evaluate solar energy harvesting in Togo, providinginsights into sustainable development.5.16.2 Small BusinessesSema-Kiosk and Cashpay are Semoa’s main products. Semoa-Kiosk which sup-port Cashpay and permit to make different virtual transactions and mainlydeposit cash on a mobile account. An innovation who permit to customer tomake transactions any time without depend of a physical place.93Eazy Chain is a digital logistics dashboard that enables small businessesto track, manage and monitor all their shipping operations by air, ocean androad. With Eazy Chain, small businesses can also pay their suppliers abroadin foreign currencies and track their consolidated shipments from all over theworld to their destinations.SocialGIS is the African Geospatial Intelligence Agency is a startup thatworks on Free Geomatics Technologies and uses big data and open data. So-cialGIS provides GIS services and solutions, and tools for data collection andvisualization, good governance, landscape restoration, nature preservation andsanitation, agriculture, urbanization and more.Dobbee Pay strives to establish a platform facilitating users in receiving andtransferring funds through mobile money, banks, and cryptocurrency platforms.This digital tool stands out by creating interoperability among existing paymentmethods, allowing users to make group payments to a maximum of one millionpeople with just one click.Solimi Fintech is on a mission to reduce cash usage by 40% within five yearsby democratizing access to financial services. The long-term vision involvesleveraging AI to create a world where all financial and commercial transactionscan be managed online. The app will feature an integrated chatbot.Artybe is a platform that seamlessly blends MI with African culture to show-case Togo as an exceptional tourist destination, emphasizing environmentalpreservation for future generations. Beyond promoting tourism, the app willfunction as a versatile learning platform covering IT, fitness, agriculture, swim-ming, and more. Users can tailor their courses based on their preferences andthe country’s wealth, considering their availability and financial capacity.5.16.3 GovernmentIn March 2021, the African Development Bank granted $2 million to the AfricanCybersecurity Resource Center (ACRC) for Financial Inclusion, aiming to com-bat cybercrime and enhance the resilience of digital financial ecosystems. Lo-cated in Lomé, the African Centre for Coordination and Research in Cybersecu-rity, established through a partnership between the government and the UnitedNations Economic Commission for Africa (UNECA), will monitor, detect, andshare cybersecurity intelligence with African governments, policymakers, lawenforcement, and security agencies. Cybercrime, costing Africa an estimated $4billion annually, remains a significant concern. The center will also spearheadinternet security research, particularly crucial as hacking groups increasinglydeploy sophisticated deep learning software to infiltrate African governmentwebsites, banks, hospitals, power companies, and telecommunication firms.Fast forward to March 2022, the inaugural Cybersecurity Summit, co-organizedby Togo and UNECA, convened Heads of State and Government, private sectorleaders, and civil society representatives to discuss Africa’s pressing cybersecu-rity challenges. During the summit, member states endorsed the Lomé Decla-ration on Cybersecurity and the Fight Against Cybercrime, commonly referredto as the Lomé Declaration. This commitment signifies member states’ pledge94to sign and ratify the African Union’s Malabo Convention, one of the world’smost comprehensive cybersecurity conventions, aiming to strengthen Africancooperation in combating cyber threats.Togo will host Artificial Intelligence Week (AIS) in March 2024. The eventwill be organized by CONIIA-Togo, the Togolese branch of the Conseil Inter-national pour l’Intelligence Artificielle, and Human-AI, a structure specializedin the development of new technologies. Scheduled from March 19-24, 2024,the SIA will gather Togo’s MI stakeholders. They will take stock of the currentstate of MI and explore its opportunities. The event will focus on raising aware-ness of advances in artificial intelligence among the general public, students,decision-makers, and institutions.Governments and humanitarian groups can use machine learning algorithmsand mobile phone data to get aid to those who need it most during a humani-tarian crisis. Researchers team helped Togo’s Ministry of Digital Economy andGiveDirectly, a nonprofit that sends cash to people living in poverty, turn thisinsight into a new type of aid program.The simple idea behind this approach, isthat wealthy people use phones differently from poor people. Their phone callsand text messages follow different patterns, and they use different data plans,for example. Machine learning algorithms - which are fancy tools for patternrecognition - can be trained to recognize those differences and infer whether agiven mobile subscriber is wealthy or poor [338].6 Eastern AfricaBurundi’s MI research landscape encompasses malaria case prediction usingdeep learning models and automated image recognition for diagnosing bananaplant diseases. In industry, there’s a focus on optimizing LPG usage withGaslink. In small businesses, Neural Labs Africa employs MI for medical im-age diagnosis, while Wiggles Technologies provides custom software solutions.The government plays a role in supporting MI adoption through initiatives likethe National MI Strategy in Seychelles. Shinzwani dictionary construction andorthographic choice is built in the Comoro Islands.In Djibouti, research spans from improving sky temperature forecasting tousing deep learning for fracture-fault detection in groundwater models. In indus-try, Farnbec addresses LPG challenges with Gaslink, and MI Connect enhancesair travel experiences. Eritrea’s research involves predictive lithologic mappingusing remote sensing data. Ethiopia applies machine learning to predict droughtand uses interpretable models for evaporation in reservoirs.In Kenya, AI Made in Africa supports startups in finding diverse tech tal-ents. The educational app, targeting school students, offers an engaging ap-proach to learning about fruits and vegetables, presenting names in French andMauritian Creole. Backed by a dataset of 1600 images, machine learning clas-sifiers were tested, revealing TensorFlow’s outstanding accuracy of 98.1%. Inthe broader context, Mauritius demonstrates a strategic approach to MI, witha national strategy and established entities like the Mauritius MI Council and95MI Academy. The government’s focus on the ocean economy aligns with thestrategy, emphasizing the potential of maritime IoT. Additionally, various com-panies, including Qubitica, Cash Radix, AgCelerant, Arie Finance, and 4SightHoldings, contribute to the integration of MI across industries, small businesses,and governmental initiatives.In Mozambique, MI activities encompass a range of research areas. Fromassessing OpenStreetMap quality using unsupervised machine learning to map-ping land use and cover, the studies reveal insights into data contributors, LULCchanges, and MI practices in education. Additionally, initiatives focus on foodsecurity, smallholder irrigated agriculture mapping, and leveraging deep learn-ing and Twitter for mapping built-up areas post-natural disasters like CycloneIdMI and Kenneth in 2019.Rwanda has approved the National MI Policy to harness MI’s benefits. InTanzania, MI applications in healthcare are explored, and Resilience Academystudents use machine learning for tree-cover mapping. Uganda focuses on creat-ing high-quality datasets for East African languages. In South Sudan, machinelearning is used to analyze fragility-related data, and in Somalia, sentimentanalysis is applied to Somali text.Meanwhile, in Zambia, machine learning aids in predicting stunting amongchildren and enhances the efficiency of health clinic verification through algo-rithms like Random Forest. The country also faces air quality challenges inmining towns, necessitating improved environmental monitoring. In Zimbabwe,the focus is on data-driven pediatrics to enhance pediatric care, effective vehicledamage classification using deep learning algorithms, and the use of technologyto predict and address adolescent depression. These initiatives underscore thetransformative potential of technology in diverse sectors across these Africannations.Each country’s MI landscape reflects a unique blend of research, industry ap-plications, small businesses, and government initiatives, showcasing the diverseways MI is contributing to development across Eastern Africa.6.1 BurundiBurundi 2020-2023 Concrete ActionsResearch ✓ waterSMB ✓InformalEconomy✓Government✓Table 37: MI in Burundi966.1.1 ResearchMalaria continues to be a major public health problem on the African conti-nent, particularly in Sub-Saharan Africa. Nonetheless, efforts are ongoing, andsignificant progress has been made. In Burundi, malaria is among the mainpublic health concerns. The work in [297] built machine-learning based mod-els to estimates malaria cases in Burundi. The forecast of malaria cases wascarried out at province level and national scale as well. Long short term mem-ory model, a type of deep learning model has been used to achieve best resultsusing climate-change related factors such as temperature, rainfall, and relativehumidity, together with malaria historical data and human population. Withthis model, the results showed that at country level different tuning of param-eters can be used in order to determine the minimum and maximum expectedmalaria cases. The univariate version of that model which learns from previousdynamics of malaria cases give more precise estimates at province-level, butboth models have same trends overall at province-level and country-level.Bananas are the dominant crop in Burundi. The surface area under cultiva-tion is estimated at 200,000 to 300,000 ha, representing 20 to 30% of the agricul-tural land. Data from Burundi’s Ministry of Agriculture and Livestock indicatefood security and nutrition continue to worsen, with 21 percent of the populationfood insecure. This could be exacerbated by various plant diseases such as theBanana Bunchy Top Disease. The disease has been reported in Angola, Benin,Burundi, Cameroon, Central African Republic, Republic of Congo, DRC, Equa-torial Guinea, Gabon, Malawi, Mozambique, Nigeria, Rwanda, South Africa,and Zambia. The East African Highlands is the zone of secondary diversity of atype of bananas called the AAA-EA types. These bananas are genetically closeto the dessert banana types but have been selected for use as beer, cooking, anddessert bananas.Banana cultivation in Burundi is grouped into three different categories.Banana for beer/wine in which juice is extracted and fermented accounts foraround 77 percent of the national production by volume. Fourteen percentof bananas are grown for cooking, and finally, about five percent are dessertbananas which are ripened and directly consumed. With recent advances inmachine learning, researchers were convinced that new disease diagnosis basedon automated image recognition was technically feasible. Minimizing the effectsof disease threats and keeping a matrix mixed landscaped of banana and non-banana canopy is a key step in managing a large number of diseases and pests[298].6.2 Comoros6.2.1 ResearchThe work in [299] discusses information and communication uses in educationthe Comoros.In [300] a Shinzwani dictionary construction and orthographic choice in theComoro Islands is presented.97Comoros 2020-2023 Concrete ActionsResearch ✓ waterSMB ✓InformalEconomy✓Government✓Table 38: MI in Comoros6.2.2 GovernmentIn 2019, the government launched the Comoros Digital Plan, which aims topromote the use of digital technologies, including AI, to drive economic growthand improve public services.In August 2023, The Banque Centrale des Comores (BCC) officially beganwork on the country’s first National Financial Inclusion Strategy (NFIS) aspart of an AFI-led training workshop. The event that occurred on 10th August,aimed to guide key stakeholders and BCC staff in formulating and implementingan NFIS to drive forward the country’s broader financial inclusion ambitions.With this move, the BCC hopes to increase access to financial services and raiseawareness among stakeholders of the pivotal role financial inclusion could playin reinforcing the country’s economic stability and the financial well-being of itspeople using machine intelligence technologies6.3 DjiboutiDjibouti 2020-2023 Concrete ActionsResearch ✓ waterSMB ✓InformalEconomy✓Government✓Table 39: MI in Djibouti6.3.1 ResearchThe building exchanges heat with different environmental elements: the sun,the outside air, the sky, and the outside surfaces. To correctly account forbuilding energy performance, radiative cooling potential, and other technicalconsiderations, it is essential to evaluate sky temperature. It is an importantparameter for the weather files used by energy building simulation software forcalculating the longwave radiation heat exchange between the exterior surface98and the sky. In the literature, there are several models to estimate sky tempera-ture. However, these models have not been completely satisfactory as far as thehot and humid climate is concerned. In this case, the sky temperature remainsoverestimated. The work in [301] is to provide a comprehensive analysis of thesky temperature measurement conducted, for the first time in Djibouti, witha pyrgeometer, a tool designed to measure longwave radiation as a componentof thermal radiation, and an artificial neural network model for improved skytemperature forecasting.A systematic comparison of known correlations for skytemperature estimation under various climatic conditions revealed their limitedaccuracy in the region, as indicated by low R2 values and high root mean squareerrors (RMSEs). To address these limitations, we introduced an ANN model,trained, validated and tested on the collected data, to capture complex patternsand relationships in the data. The ANN model demonstrated superior perfor-mance over existing empirical correlations, providing more accurate and reliablesky temperature predictions for Djibouti’s hot and humid climate. This studyshowcases the effectiveness of an integrated approach using pyrgeometer-basedsky temperature measurements and ANNs for sky temperature forecasting inDjibouti. Our findings support the use of advanced machine learning techniquesto overcome the limitations of existing correlations and improve the accuracy ofsky temperature predictions, particularly in hot and humid climates.The work in [302] examines fracture-fault detection using deep learning. Ac-curate estimation of groundwater flow is crucial in arid regions where permanentsurface water is absent. In several groundwater simulation models, an importantparameter for identifying areas with high potential for groundwater resourcesis the accurate fracture-fault detection. In the present study we propose adeep learning approach to detect fracture-fault structures in the Ali Faren sub-catchment of Ambouli Wadi in Djibouti. Our deep convolutional neural network(Deep-CNN) model is trained on high-spatial resolution multispectral satelliteimages using wadi streamline as labels. Fracture-fault structures are extractedusing stepwise elimination based on geological characteristics observed in re-lief images derived from PALSAR-1/2 data. Their results demonstrate thatthe proposed Deep-CNN model accurately detects fracture-fault lines, achiev-ing a validation accuracy of 0.9684, precision of 0.9124, recall of 0.9701, andF1 of 0.8997. The proposed model has the potential to identify potential areasfor groundwater resources across the country, contributing to sustainable watermanagement and improving Djibouti’s water security.6.4 Eritrea6.4.1 ResearchA regional bedrock map provides a foundation from which to build geologicalinterpretations. However, rapid and accurate bedrock mapping in an area thatlacks outcrop is a common problem, especially in regions with sparse data. Ahistoric bedrock map from an Au and base metal project in the Kerkasha dis-trict, Eritrea, is significantly improved by predicting bedrock distribution in99Eritrea 2020-2023 Concrete ActionsResearch ✓ waterSMB ✓InformalEconomy✓Government✓Table 40: MI in Eritreaareas previously mapped as transported overburden. Publicly-available remotesensing data (DTM and ASTER) were combined with airborne geophysical data(magnetics and radiometrics) to provide features for bedrock prediction [303]. Remote sensing data were pre-processed using Principal Components Analy-sis to yield an equal number of principal components as input features. Fouriterations were trialled, using different combinations of remote sensing PC fea-tures. The two initial trials used all available remote sensing data but comparedresults when feature ranking and selection is applied to reduce the number ofPCs used for training and classification. The subsequent two trials used sub-sets of available remote-sensing data, selected based on domain expertise (i.e.,the domain-specific knowledge of a geologist), with all respective PCs were re-tained. Five-fold cross-validation scores were highest when a DTM, magnetics,and radiometrics data were included as input features. However, qualitative vi-sual appraisal of predicted results across trials, complemented by maps of classmembership uncertainty (using a measure of entropy), indicate that geologically-meaningful results are also produced when radiometrics are omitted and onlythe DTM and magnetics are used. The study concludes with a generalized work-flow to assist geologists who are seeking to improve the bedrock interpretationof areas under cover in a single area of interest. Domain expertise is shown to becritical for the selection of appropriate input features and validation of resultsduring predictive lithologic mapping.6.5 EthiopiaEthiopia 2020-2023 Concrete ActionsResearch ✓ waterSMB ✓ iCog LabsGovernment✓ EAIITable 41: MI in Ethiopia1006.5.1 ResearchThis study [304] applies machine learning to the rapidly growing societal prob-lem of drought. Severe drought exists in Ethiopia with crop failures affectingabout 90 million people. The Ethiopian famine of 1983-85 caused a loss of400,000 - 1,000,000 lives. The present drought was triggered by low precip-itation associated with the current El Niño and long-term warming, enhanc-ing the potential for a catastrophe. In this study, the roles of temperature,precipitation and El Niño are examined to characterize both the current andprevious droughts. Variable selection, using genetic algorithms with 10-foldcross-validation, was used to reduce a large number of potential predictors (27)to a manageable set (7). Variables present in 70% of the folds were retained toclassify drought (no drought). Logistic regression and Primal Estimated sub-GrAdient SOlver for SVM (Pegasos) using both hinge and log cost functions,were used to classify drought. Logistic regression (Pegasos) produced correctclassifications for 81.14% (83.44%) of the years tested. The variable weightssuggest that El Niño plays an important role but, since the region has under-gone a steady warming trend of 1.6 Celsius since the 1950s, the larger weightsassociated with positive temperature anomalies are critical for correct classifi-cation.The work [305] develops an Interpretable machine learning for predictingevaporation from Awash reservoirs in Ethiopia. An in-depth understanding ofa key element such as lake evaporation is particularly beneficial in developingthe optimal management approach for reservoirs. In this study, we first aimto evaluate the applicability of regressors Random Forest, Gradient Booting,and Decision Tree, K-Nearest Neighbor, and XGBoost architectures to predictdaily lake evaporation of five reservoirs in the Awash River basin, Ethiopia.The best performing models, Gradient Boosting and XGBoost, are then ex-plained through an explanatory framework using daily climate datasets. Theinterpretability of the models was evaluated using the Shapley Additive ex-planations (SHAP). The factors with the greatest overall impact on the dailyevaporation for GB and XGboost Architecture were the SH, month, Tmax, andTmin for Metehara and Melkasa, and Tmax, Tmin, and month had the greatestimpact on the daily evaporation for Dubti. Furthermore, the interpretabilityof the models showed good agreement between the simulations and the actualhydro-climatic evaporation process. This result allows decision makers to notonly rely on the results of an algorithm, but to make more informed decisionsby using interpretable results for better control of the basin reservoir operatingrules.6.5.2 Small BusinessesiCog Labs is a team of software professionals dedicated to advancing the fron-tier of research and applications in machine intelligence and delivering qualityproducts to clients. It is based in Addis Ababa, Ethiopia.1016.5.3 GovernmentThe Artificial Intelligence and Robotics is one of the centers of excellence whichis identified by the ministry of science and technology to be established in AddisAbaba Science and Technology University.The Artificial Intelligence & Roboticscenter of excellence (AI&R CoEs) is established with the aim to create a closecollaboration between the academia and industries in the fields of Artificialintelligence and robotics.The Ethiopian Artificial Intelligence Institute (EAII) has become African Ar-tificial Intelligence Center of Excellence, the Ministry of Innovation and Tech-nology of Ethiopia (MInT) confirmed in November 2023. The Ethiopian AIinstitute was promoted as the continent’s center of excellence during (5th Or-dinary Session of the African Union Specialized Technical Committee on Com-munication and ICT (STC-CICT-5) which is being held at the African Unionconference hall, in Addis Ababa. The Ethiopian Artificial Intelligence Institutewas proposed to become the “African Artificial Intelligence Center of Excel-lence”. The proposal was accepted and approved by the members of the ICTand Communication Ministers of African countries.6.6 KenyaKenya 2020-2023 Concrete ActionsSMB ✓ NeuralSight, AIfluence,Amini, Halkin, Freshee, M-Shule, AI ConnectInformalEconomy✓Government✓ 2019 Kenya’s Dis-tributed Ledger Tech-nology and ArtificialIntelligence Taskforce ,AICEATable 42: MI in Kenya6.6.1 ResearchYego et al. (2021) conducted a comparative analysis of machine learning modelsfor predicting insurance uptake in Kenya, emphasizing the role of insurance infinancial inclusion and economic growth [149]. Mulungu et al. present a machinelearning approach to assess the economic impact of integrated pest managementpractices for mango fruit flies in Kenya [150]. Alharahsheh and Abdullah (2021)predict individuals’ mental health status in Kenya using machine learning meth-ods [151]. Yego et al. (2023) optimize pension participation in Kenya through102a comparative analysis of tree-based machine learning algorithms and logisticregression classifier [152]. Pius et al. (2021) employ supervised machine learningto model the demand for outpatient health-care services in Kenya using artificialneural networks and regression decision trees [153]. Shah et al. (2023) predictpostpartum hemorrhage (PPH) in a Kenyan population using machine learningalgorithms [154]. Ondiek et al. (2023) develop a recommender system for STEMenrollment in Kenyan universities using machine learning algorithms [155]. Wil-son et al. (2017) demonstrate that ensemble machine learning and forecastingcan achieve 99% uptime for rural handpumps [156]. Lees et al. (2022) applydeep learning for vegetation health forecasting in Kenya [157]. Gram-Hansenet al. (2019) map informal settlements in developing countries using machinelearning and low-resolution multi-spectral data [158]. Orare (2019) developsa travel time prediction model for Nairobi city using machine learning algo-rithms [159]. Kochulem et al. (2023) conduct a mass valuation of unimprovedland value in Nairobi County [160]. Pius Kamando (2023) proposes a tree-basedneural network for forecasting outpatient health-care services demand in NairobiCounty, Kenya [161]. Kuria (2014) utilizes machine learning for flood forecast-ing in the Nzoia river basin, western Kenya [162]. Onyango (2021) develops aTwitter sentiment analysis tool for detecting crime hotspots in Nairobi, Kenya[163]. Magiya (2020) predicts package delivery time for motorcycles in Nairobi[164]. Muthoka et al. (2021) map Opuntia stricta in the arid and semi-aridenvironment of Kenya using Sentinel-2 imagery and ensemble machine learn-ing classifiers [165]. Omolo (2016) creates a mobile and web-based applicationfor security intelligence gathering in Nairobi County [166]. Mbani et al.(2020)employ artificial intelligent agents for crime mapping in Nairobi City County,Kenya [167]. Omondi and Boitt (2020) model the spatial distribution of soilheavy metals using a random forest model- a case study of Nairobi and ThirirkaRivers’ confluence [168]. Muchuku (2023) assesses recurrent neural networksas a prediction tool for quoted stock prices on the Nairobi Securities Exchange[169].6.6.2 Small BusinessesHalkin designs, manufactures and operates Unmanned Aerial Systems (UAS).Halkin is able to implement embedded systems through our software engineersfor additional capability such as faster processing; incorporation to various sen-sors and systems, providing our own failsafe procedures; Incorporation of imageprocessing, MI and Machine Learning.Freshee is a mobile marketplace for deals at every food, drink and enter-tainment venue to help users Save More and Explore while getting rewarded forloyalty at their favorite places. The venue discovery industry is broken. Venuesface several hours of low/no footfall daily, struggle to advertise promotional of-fers, and effectively subscribe customers to loyalty programs. Customers missout on deals and venues they would love to visit.Neural Labs Africa is an innovative medical technology Company using Ar-tificial Intelligence to transform medical imaging diagnosis. We have devel-103oped (NeuralSight) a technology that screens medical images for Radiologistsand Hospitals in real-time. NeuralSight can identify over 20 respiratory, heartand breast diseases which include: Pneumonia, Tuberculosis, COVID-19, Pneu-mothorax, Cardiomegaly, Benign breast Tumor, Malignant breast Cancer, At-electasis, Infiltration, etc.AI connect is on Conversational MI and Omni-channel customer engage-ment platform that connects the air traveler to the airline ecosystem usingartificial intelligence, machine learning, and customer engagement excellence.Farnbec adresses firsthand the challenges and inconveniences of cooking withLPG. That’s why Farnbec developed Gaslink as a solution. Gaslink, is beingdeveloped with the goal of revolutionizing the way that Households and Restau-rant Chains manage their LPG usage for clean cooking. Using advanced tech-nology including NB-IoT, cloud computing, AI, and APIs, our solution providesreal-time tracking and monitoring of LPG cylinder usage, as well as valuableinsights and recommendations.Wiggles Technologies is a custom software development company that pro-vides dedicated groups of highly-skilled and creative programmers. We delivercustom software applications and mobile solutions, run software testing, per-form in-depth product analyses, and provide technology management, supportand expertise.AI Made in Africa helps Startups and SMEs find talented, diverse tech tal-ents by matching them with candidates of the best culture fit while providingpractical levels of flexibility.Founded in 2017, M-Shule is the first personalized knowledge-building plat-form in Africa, connecting learners to tailored learning, evaluation, activation,and data tools through SMS and chatbot. Meaning ”mobile school” in Swahili,M-Shule combines SMS with artificial intelligence to reach offline or marginal-ized communities, offering self-paced, interactive, and personalized resources.Initially focusing on academic courses, M-Shule has expanded to include pro-fessional courses, life skills, data collection, and behavior change. To date, theplatform has reached over twenty thousand households, not only in Kenya butalso across East Africa. M-Shule has demonstrated success in over 30 Kenyancounties, Uganda, and Tanzania, covering more than 6 skill development do-mains and 7 languages, including Dholuo, English, Kamba, Kikuyu, Kiswahili,Ng’aturkana, and Somali.Nairobi-based climate tech startup Amini is focused on solving Africa’s en-vironmental data gap through artificial intelligence and satellite technology andhas raised $2 million in a pre-seed funding round. The Kenyan startup wasfounded in 2022 was designed to address Africa’s data scarcity, facilitate capitalinvestment, promote climate resilience, and accelerate economic developmentopportunities in the region. Furthermore, the platform also provides access tovaluable environmental data analytics, including drought, flood, soil and crophealth. This data can be processed to forecast crop yields for smallholder farm-ers in seconds and to measure the impact of natural disasters across the region.Before the funding, the company initially focused on the insurance industry,however, it is now experiencing rapid expansion into supply chain monitoring,104specifically at the ”last mile”, or the initial stages of the global supply chain.Founded in 2019, AIfluence uses advanced machine learning algorithms tomatch influencers with a target demographic through its audience-first strategy.The Kenyan startup in 2021 raised a $1 million seed funding round to accel-erate the expansion of its MI-powered marketing platform. The MI-poweredmarketing platform allows advertisers to onboard and coordinate hundreds tothousands of micro and nano influencers per campaign, generating authenticpeer-to-peer conversations and superior conversion. Sky.Garden is a Kenyanmobile SaaS eCommerce Platform for African retailers.6.6.3 GovernmentIn 2019, Kenya’s Distributed Ledger Technology and Artificial Intelligence Task-force report provided the government with a strategic direction on developinga roadmap to uphold human rights when adopting emerging technologies likeMI. The report recommends leveraging blockchain and MI to combat corrup-tion and enhance state transparency. The report assesses emerging technolo-gies and their deployment globally, recommending that the government utilizeBlockchain Technology and machine intelligence to combat and eliminate cor-ruption, safeguarding the interests of citizens. It advises the government toleverage Blockchain and MI technology solutions to fight corruption and en-hance transparency due to their record immutability.AICE: Founded in 2020, the AI Centre of Excellence is passionate about cre-ating value and sustainable impact within the African Intelligence and MachineLearning space by Transforming Data Scientists & Software engineers into MIand ML Engineers, Creating sustainable impact through Research and Devel-opment, Providing custom MI as a Service and building MI solutions. Fromchallenges to solutions, the AI Centre of excellence aims to develop impactwithin the MI space that allows for growth, innovation and creativity.6.7 MadagascarMadagascar 2020-2023 Concrete ActionsResearch ✓ waterSMB ✓InformalEconomy✓Government✓Table 43: MI in Madagascar1056.7.1 ResearchThe work in [183] by Clément Le Ludec, Maxime Cornet, Antonio A Casilliexplores the impact of MI on labor, focusing on France outsourcing tasks toworkers in Madagascar. The study unveils the intricate production chain ofMI, revealing the reliance on data workers in low-income countries. The workin [184] by Fahafahantsoa Rapelanoro Rabenja discusses the PASSION Projectin Madagascar and Guinea, using MI for dermatological data collection. Thestudy aims to address the scarcity of dermatologists, emphasizing the potentialof MI in enhancing data collection on skin conditions. The work in [185] byPaola Tubaro, Antonio A Casilli, Marion Coville delves into the role of dig-ital platform labor in MI development. The study highlights micro-workers’functions in MI preparation, verification, and impersonation, emphasizing theenduring significance of micro-work in contemporary MI production processes.The work in [186] by Daniele Silvestro, Stefano Goria, Thomas Sterner, Alexan-dre Antonelli introduces a framework, CAPTAIN, for spatial conservation pri-oritization using reinforcement learning. The study demonstrates the efficacyof MI in maximizing biodiversity protection under limited budgets, presentinga promising approach for conservation in a resource-limited world. The workin [187] by Sandro Valerio Silva, Tobias Andermann, Alexander Zizka, GregorKozlowski, Daniele Silvestro addresses the global conservation crisis for treespecies. The study employs MI to estimate and map the conservation status ofover 21,000 tree species, revealing insights into threatened species distributionand providing efficient approximations of extinction risk assessments.The work in [188] by Harimino Andriamalala Rajaonarisoa et al. charac-terizes the evolution of precipitation in Southern Madagascar using High OrderFuzzy Time Series. The study models annual precipitation data, determin-ing hyperparameters and fuzzy sets to interpret the characteristic evolution ofprecipitation. The work in [189] by RABENIAINA Anjara Davio Ulrick andRAKOTOVAO Niry Arinavalona presents a method for modeling the onset andend dates of the monsoon season in Northern Madagascar. It employs a Ma-chine Neural Fuzzy Inference System (ANFIS) based on MI and zonal winddata, providing estimates for the monsoon season. The work in [190] by JBKoto, TR Ramahefy, S Randrianja focuses on the extraction of knowledge fromcivil status data (surname and first name) using MI. The study demonstratesthe application of MI and Python tools to analyze and visualize patterns in theformulation of names in Madagascar. The work in [191] by Paola Tubaro andAntonio A Casilli explores the role of micro-work in the ”back-office” of MI, par-ticularly in the automotive industry. The study highlights the labor-intensiveprocess of MI production, emphasizing the structural need for micro-workers indata annotation, tagging, and labeling for smart solutions in the industry.The work in [192] by TR Rasamoela and J Szpytko explores the implemen-tation of telematics in the transport system in Antananarivo, Madagascar. Thepaper emphasizes the significance of reliable transportation for economic growthand poverty reduction, proposing the use of telematics as a solution to enhancethe transport sector in Madagascar. The work in [193] by Manuel Dominguez-106Rodrigo et al. introduces a breakthrough method that utilizes MI and computervision techniques to achieve high accuracy in the classification of modern andancient bone surface modifications. The study demonstrates the potential of MIin objectively identifying hominin butchery traces in the archaeological record.The work in [194] by Matteo Giuliani et al. presents the Climate State Intel-ligence framework, employing MI to detect the state of multiple global climatesignals. The framework enhances seasonal forecasts, particularly in the LakeComo basin, providing valuable information for water system operations andimproving system performance. The work in [195] by Dominique Badariotti etal. introduces SIMPEST, an agent-based model designed to simulate plague epi-demics in Madagascar. The research focuses on understanding the behavior andspread of plague in the environment, aiming for better control and managementof this epidemiological case. The work in [196] by Ala Saleh Alluhaidan investi-gates public perception of drones as a tool for telecommunication technologies.The study explores how the public views drones, particularly in healthcare ap-plications, and identifies concerns related to safety, security, and privacy. Theresults highlight the need for increased public awareness and education aboutdrone technology.The work in [197] examines the use of fuzzy inference modeling to predictthe beginning and ending dates of rain in the coastal areas of South East Mada-gascar. The model, based on MI and fuzzy logic, covers the period from 1980to 2017 and demonstrates excellent performance with a calculated MAPE ofless than 10%. The work in [198] presents an abstractive text summarizationapproach for the Malagasy language. Utilizing the Scheduled Sampling modeland deep learning, the study focuses on summarizing content in a more naturaland harmonious manner. The results indicate the applicability of deep learningto the Malagasy language. The work in [199] focuses on estimating deforesta-tion in tropical humid and dry forests in Madagascar from 2000 to 2010. Usingmulti-date Landsat satellite images and a random forests classifier, the studyprovides high-resolution deforestation maps with reliable uncertainty estimates,crucial for forest conservation and management. The work in [200] exploresthe enhancement of a budget simulation model for decentralized territorial au-thorities in Madagascar using MI. The study emphasizes the importance ofpredictive analyses in better managing budget implementation by consideringvarious factors such as economic, political, and performance indicators. Thework in [201] investigates climatic factors affecting monthly rainfall variabilityin a remote region of Madagascar. Machine learning models, analyzing pastweather conditions and relevant climate indices, contribute to the developmentof short-to-medium-range rainfall outlook models. The work in [202] appliesan Machine neural network approach to forecast infant mortality rate in Mada-gascar. Covering the period 1960-2020, the study’s stable model predicts thatthe infant mortality rate will be around 35/1000 live births per year in the out-of-sample period, emphasizing the need for maternal and child care programs.The work in [203] introduces a one-dimensional convolutional neural network forvisible and near-infrared spectroscopy to improve soil phosphorus prediction inMadagascar. The study demonstrates the model’s superior predictive accuracy107compared to traditional regression methods, contributing to effective fertilizermanagement and ecosystem sustainability.6.8 MalawiMalawi 2020-2023 Concrete ActionsResearch ✓ waterSMB ✓InformalEconomy✓Government✓ Centre for Artificial In-telligence and STEAM- Science, Technology,Engineering, Arts andMathematicsTable 44: MI in Malawi6.8.1 ResearchIn [204], Poverty alleviation in Malawi is explored through machine learningmodels utilizing existing survey data to predict poor and non-poor households.Open-source algorithms such as Logistic Regression, Extra Gradient BoostingMachine, and Light Gradient Boosting Machine demonstrate accuracy compa-rable to full feature sets, suggesting the potential for shorter, lower-cost surveys.In [205] machine learning is used, specifically a random forest model, onhigh-frequency household survey data in southern Malawi to infer predictors offood insecurity. The model outperforms others, emphasizing the significance oflocation and self-reported welfare as predictors. Various models are evaluatedfor forecasting food security outcomes.[206] introduces an energy-climate-water framework, combining satellite dataand machine learning, to assess the impact of hydro-climatic variability on hy-dropower reliability in Malawi. The approach, validated for the period 2000 -2018, mitigates data scarcity and enhances understanding of vulnerabilities inthe power sector.The work in [207] examines legislation in Malawi. Legal research in Malawifaces challenges with limited resources. This interdisciplinary research buildstools for annotating Malawi criminal law decisions with legal meta-data usingmachine learning tools, spaCy, and Gensim LDA. The study sets the foundationfor classifying Malawi criminal case law according to the International Classifi-cation of Crime for Statistical Purposes.In [208], the authors compare Machine Learning methods with hedonic pric-ing using household survey data from Uganda, Tanzania, and Malawi. ML108methods such as Boosting, Bagging, Forest, Ridge, and LASSO outperformOLS models, providing superior prediction of rental values in housing surveys.The work in [209] explores the adoption of conservation agriculture in Malawi,finding that peer effects, particularly adoption by neighbors, play a crucial role.The study highlights the significance of considering social dynamics and peerinfluence in promoting CA interventions.6.8.2 GovernmentIn October 2023, Malawi launched its first-ever Centre for Artificial Intelligenceand STEAM - Science, Technology, Engineering, Arts and Mathematics - atthe Malawi University of Science and Technology. Established with supportfrom various U.S.-based universities, the center aims to provide solutions to thecountry’s innovation and technology needs.6.9 MauritiusMauritius 2020-2023 Concrete ActionsResearch ✓ FruVegySMB ✓ Qubitica, Cash Radix,AgCelerant, Arie Fi-nance, 4SightTable 45: MI in Mauritius6.9.1 ResearchThe study [210] looks at plants called invasive flora alien species (IAS), whichcan harm the variety of life in tropical forests. They focused on one specificplant, strawberry guava, and used pictures from satellites that anyone can accessto learn more about how these plants affect tropical forests. This might bethe first time someone used these free satellite pictures to create a map ofstrawberry guava and the first time they used this method to map invasivespecies in Mauritius.In a park in Mauritius called Black River Gorges National Park (BRGNP),the researchers did some on-the-ground observations and collected 4670 samplesto understand how much strawberry guava covered different areas. They used70% of this information to teach their computer models and make them better,and the other 30% they kept to test how accurate their models were. They usedspecial satellite images and a tool called Google Earth Engine for this. Theyalso used some calculations to help them understand the colors and textures ofthe strawberry guava plants in the pictures.Their computer models, called Random Forest and Support Vector Machine,did a really good job. RF was recommended for future studies because it was109very accurate (97.60% ± 0.20% with 95% confidence) and made predictionsin more consistent areas. They also found that strawberry guava was mostcommon in the central parts of BRGNP and on steeper slopes. Surprisingly, theamount of strawberry guava didn’t change much from 2016 to 2020.In [211], the authors explore new ways, like using computers to analyze lotsof data, to predict how bad accidents might be.They tried different computer methods, like Support Vector Machine, Gra-dient Boosting, Logistic Regression, Random Forest, and Naive Bayes, all usinga programming language called Python. The method called Gradient Boostingdid the best job in figuring out how severe accidents could be. It was rightabout 83.2% of the time, which is pretty good, and it had an AUC of 83.9%,showing it’s effective in making these predictions.[212] examines proper identification of plant species has major benefits fora wide range of stakeholders ranging from forestry services, botanists, tax-onomists, physicians, pharmaceutical laboratories, organisations fighting for en-dangered species, government and the public at large. Consequently, this hasfueled an interest in developing automated systems for the recognition of differ-ent plant species. A fully automated method for the recognition of medicinalplants using computer vision and machine learning techniques has been pre-sented. Leaves from 24 different medicinal plant species were collected andphotographed using a smartphone in a laboratory setting. A large number offeatures were extracted from each leaf such as its length, width, perimeter, area,number of vertices, colour, perimeter and area of hull. Several derived featureswere then computed from these attributes. The best results were obtained froma random forest classifier using a 10-fold crossvalidation technique. With anaccuracy of 90.1%, the random forest classifier performed better than other ma-chine learning approaches such as the k-nearest neighbour, naive Bayes, supportvector machines and neural networks. These results are very encouraging andfuture work will be geared towards using a larger dataset and high-performancecomputing facilities to investigate the performance of deep learning neural net-works to identify medicinal plants used in primary health care. To the bestof our knowledge, this work is the first of its kind to have created a uniqueimage dataset for medicinal plants that are available on the island of Mauri-tius. It is anticipated that a web-based or mobile computer system for theautomatic recognition of medicinal plants will help the local population to im-prove their knowledge on medicinal plants, help taxonomists to develop moreefficient species identification techniques and will also contribute significantly inthe protection of endangered species.The research work [213] studies flood prediction using Machine neural net-works in Mauritius. The average temperature of the earth is increasing at analarming rate and it has been envisaged to increase by a factor of about 1.4 to5.8 degree Celsius by the year 2100. An increase in the atmospheric tempera-ture entails the occurrence of many extreme events such as stronger heat waves,formation of intense cyclones, unprecedented flash floods and severe droughtevents which are set to impact greatly on both the global economy and society.Among the various natural disasters, which affect mankind, flash floods have110been reported to cause more casualties in terms of economic loss, death tolls andinfrastructural damages. Flooding has become a recurrent phenomenon in therecent decade accounting for about 73% of damages caused by natural disasterswhich in turn results in an overall loss of about $30 billions. Flash floods arethus a global phenomenon affecting major parts of the world as indicated for theyear 2018, which marked the occurrence of several deadly flash floods in Kerala,France and Vietnam. In [213] the focus is on Mauritius, which is a small islandlocated in the Indian Ocean, off the east coast of Africa and Madagascar. Themorphological landscape of Mauritius consists of highlands and coastal regionsin a relatively small geographical area of 1865 km² such that it is typical for theisland to experience several microclimates on the same day in different regions.Their study is especially motivated by the occurrence of a series of flash floodsin Mauritius.Receiving and managing complaints effectively are important for organisa-tions which aim to provide excellent customer service. In order for this tohappen, organisations should make it quick and easy for users to report issues.In [214] , a smart mobile application for complaints management in Mauritiusis described. Users of this mobile application can report issues for differentorganisations using a single application on their smartphones. They can reg-ister complaints using text, images or videos, and they do not have to specifywhich authority the complaint is directed to. Instead, the application uses textand image analysis alongside a Convolutional Neural Network in order to di-rect complaints to the correct utility organisations. The classifiers have beentrained to identify different categories of complaints for each local utility organ-isation. Users are notified regarding the status of their complaints and can usethe application to directly communicate with the personnel.Small Island Developing States (SIDS), like Mauritius, share similar sus-tainable development challenges inherent to their characteristics. Growth inthe global energy demand and fears of energy supply disruptions, have trig-gered much debate geared towards the necessity for sustainable energy plan-ning. Accurate forecasting of future electricity demand is an essential input tothis process. Such forecasts are also important in regional or national powersystem strategy Management. Non linearity of the factors adds complexity tothe electricity load forecasting process. Statistical learning theory, in the formof Support Vector Machines, have been used successfully to tackle nonlinearregression and time series problems. However application to the electricity de-mand forecasting problem with focus on SIDS’characteristics is lacking. Thearticle [215] focuses on the application of SVMs to forecast electricity demandof a SIDS member, Mauritius. A two years ahead forecast, for 2008 and 2009,was derived using monthly time series data from years 1996 to 2007. The inputsconsidered were historical electricity demand and prices, temperature, humidity,population and GDP.To facilitate the recognition and classification of medicinal plants that arecommonly used by Mauritians, a mobile application which can recognize sev-enty different medicinal plants has been developed in [216]. A convolutionalneural network based on the TensorFlow framework has been used to create the111classification model. The system has a recognition accuracy of more than 90%.Once the plant is recognized, a number of useful information is displayed tothe user. Such information includes the common name of the plant, its Englishname and also its scientific name. The plant is also classified as either exoticor endemic followed by its medicinal applications and a short description. Con-trary to similar systems, the application does not require an internet connectionto work. Also, there are no pre-processing steps, and the images can be takenin broad daylight. Furthermore, any part of the plant can be photographed.It is a fast and non-intrusive method to identify medicinal plants. This mobileapplication will help the Mauritian population to increase their familiarity ofmedicinal plants, help taxonomists to experiment with new ways of identifyingplant species, and will also contribute to the protection of endangered plantspecies.The research article in [217] investigates the application of supervised ma-chine learning techniques to predict the price of used cars in Mauritius. Thepredictions are based on historical data collected from daily newspapers. Differ-ent techniques like multiple linear regression analysis, k-nearest neighbors, naivebayes and decision trees have been used to make the predictions. The predic-tions are then evaluated and compared in order to find those which provide thebest performances. A seemingly easy problem turned out to be indeed very dif-ficult to resolve with high accuracy. All the four methods provided comparableperformance.In [218] a Machine Learning Technique called the Support Vector Machineis adopted on the Stock Exchange of Mauritius (SEM) to determine if stockmarket returns are predictable based on information from past prices, allowingarbitrage opportunities for abnormal profit generation. The serial correlationtest, used as benchmark, and the SVM technique show evidence that previousinformation on share prices as well as the indicators constructed are usefulin predicting share price movements. The implications of the study are thatinvestors have the prospect of adopting speculative strategies and profits fromtrading based on information and advanced techniques and models are possible.In this era of education and technology, it is undeniable that there is a grow-ing interaction between machine and humans. Student performance is of primeimportance as education is the key to success. At the university of Mauritius,the number of students enrolled in a course does not match the number of stu-dents graduating as not every student complete their academic cycle of 3 or4 years. Some extend their course duration as they have to repeat the wholeyear or several modules, while others exit with a certificate or diploma sincethey lack the required number of credits to obtain a degree. Unfortunately, theregistration of some students with very low average marks are terminated. Theresearch work [219] investigates a machine learning model to predict the perfor-mance of university students on a yearly basis. The model will forecast studentperformance and help take necessary actions before it is too late. The classifi-cation technique is used to train the proposed model using an existing studentdataset. The training phase generates a training model that can then be usedto predict student performance based on parameters such as attendance, marks,112study hours, health or average performance. Different algorithms are evaluatedand the classification and prediction algorithms which are more accurate arerecommendedThe Mauritius MI Strategy 2018, established by the government, aims atmaking MI a cornerstone of the next development model by recognizing thepotential of technology to improve growth, productivity and quality of life. Inthis regard, MI has already started to shape the legal sector, for instance, byassisting law practitioners to identify and minimize bias in client intake, offerinitial consultation solutions, expand the scope of information for law practition-ers and predict the outcome of future legal cases, among others. Nevertheless,while the legal profession worldwide is facing pressure to innovate and trans-form, the emergence of MI is causing significant disruption to long-establishedpractices in the legal world, especially since this particular sector has tradition-ally under-utilized technology. The work in [220] seeks to assess the influence ofMI on employees from the legal profession mainly in terms of their performance,their reaction, and adaptability to change and to identify the challenges facedby these employees in Mauritius in adopting MI for their operational activities.Sentiment analysis is becoming increasing important with the rise in theamount of content on social media. However, sentiment analysis remains chal-lenging for under-resourced languages such as Kreol Morisien (KM), the nativelanguage of Mauritius. In fact, it has been observed that in Mauritius, socialmedia comments often consist of more than one language among English, Frenchand Kreol Morisien. In work in [221] first creates an annotated dataset of 1300sentences and then outline a framework through which sentiment analysis canbe performed on social media comments. We propose a KM sentiment analyzerusing two algorithms namely Support Vector Machine (SVM) and MultinomialNaive Bayes (MNB). Our results show that SVM outperforms MNB for sen-timent analysis in Kreol Morisien, achieving an accuracy of 66.15% after pre-processing techniques stopwords removal and spell checking are applied. Thispaper highlights the need to develop further tools in order to enable naturallanguage processing of Kreol Morisien.Fruit Flies impact the field of agriculture in a negative way affecting theeconomy of a host country. This work in [222] presents an identification systemthat can be deployed on a mobile application. The identification system usesConvolutional Neural Network (CNN) to learn the key visual features of fruitflies to be able to perform detections. The system comprises of two main as-pects; a detection model and a classification model. The detection model usesSingle Snapshot Detector (SSD) MobileNet V2 FPNLite 640x640 model which isconverted to a TensorFlow Lite version and hosted locally on the mobile phone.The classification model uses Xception model which is hosted on the GoogleCloud Platform where requests are made to the cloud from the mobile applica-tion. A custom image dataset of nine(9) fruit flies was created in two ways: Twofruit flies predominant in Mauritius were obtained from the Entomology Depart-ment of the Ministry of Agro Industry and Food Security and photographs weretaken. The remaining fruit flies’ images were obtained through web scraping.Transfer learning has been successfully used to produce the SSD MobileNetV2113FPNLite 640x640 model with a loss of 31% and the Xception model with anaccuracy of 75.5%.Birds communicate with their colonies through sound and inform them ofpotential problems like forest fires. The identification of bird sounds is thereforevery important and has the potential to solve some global problems. Convolu-tional neural networks (CNNs) are sophisticated deep learning algorithms thathave proven to be effective in image processing and in sound classification. Thework in [223] describes the work done to develop a tool using a deep learningmodel for classifying Mauritius bird sounds from audio recordings. A datasetobtained from the Xeno-canto bird song sharing site, which hosts a vast collec-tion of labeled and classified recordings, is used to fine-tune three pre-trainedCNN models, namely InceptionV3, MobileNetV2 and RestNet50 and a custommodel. The neural network’s input is represented by spectrograms created fromdownloaded mp3 files. Time shifting and pitch stretching have been used fordata augmentation. The best performing model has been integrated into a web-site to identify birds sounds recordings. In this work, transfer learning has beenused successfully to produce a model with a weighted accuracy of 84%. Al-though a custom CNN was trained, better accuracy was achieved through theuse of transfer learning.Nowadays, many people are unaware of the benefits of fruits and vegetableswhich has resulted in their reduced consumption. This has inevitably led toa rise in diseases such as obesity, high blood pressure and heart diseases. Tothis end, [224] developed FruVegy which is an android app which can automat-ically identify fruits and vegetables and then display its nutritional values. Theapp can identify forty different fruits/vegetables. The app is specially targetingschool students who will find it easy and fun to use and this, we believe, willincrease their interest in the consumption of fruits and vegetables. Furthermore,the names of the fruits and vegetables are also available in French and in Mau-ritian Creole. A dataset of 1600 images from 40 different fruits and vegetablesis proposed. There was an equal number of images for each fruit/vegetable.Features such as shape, color and texture were extracted from each image. Dif-ferent machine learning classifiers were tested but random forest with 100 treesproduced the best result with an accuracy of 90.6%. However, with TensorFlow,an average accuracy of 98.1% was obtained under different scenarios.6.9.2 Small BusinessesQubitica: Decentralized autonomous organization. It offers QBIT, an ERC-20token that serves as a membership token and a currency. It works on blockchainand MI projects with the help of developers and entrepreneurs. It allows in-vestors to invest in projects these projects. It focuses on projects such as mining,trading & exchange, news platform, voting system, node services, accounting,and more.Cash Radix: AI-based trading signals for forex market assets. It providesforex recommendations with the market analysis, Machine intelligent technicalanalysis software robot, indicators, and more. It features charts, fundamental114analysis, broker recommendation, forex, calculator, and more. It also offers atraining platform for forex market trading.AgCelerant: Provider of phygital agriculture solutions to agribusinesses us-ing earth observation, IoT, and MI. They offer physically based, digitally-drivensolutions to secure the sustainability, transparency, and sourcing of food. Theplatform connects smallholders contract farming, producers with banks, insur-ers, input providers, and agro-industries to control risks, secure transactions,reduce frictions and sustainably improve productivity and welfare. It imple-ments a business model in which smallholder farmers and larger-scale investorsare simultaneously accompanied and protected as they empower themselves torespond to growing and changing customer needs.Arie Finance : Software for core banking. It offers core banking solutionsthat enable bank accounts, cashflows, payments, investments, and more. Itprovides a dashboard to monitor verifications, customers, exchange rates; om-nichannel communication including mobile and web applications, and more.4Sight Holdings Limited (4Sight) is a public company listed on the JSE AltXincorporated on 29 June 2017 in accordance with the laws of the Republic ofMauritius. 4Sight focuses on a cross-section of established, new, and emerg-ing technologies. These include MI solutions with machine learning, big data,cloud and business intelligence solutions, digital twin and simulation, informa-tion and operational technologies, production scheduling, horizontal and verticalintegration, industrial internet of things, cloud service provider, robotic processautomation and augmented and virtual reality solutions.6.9.3 GovernmentIn 2018, the government has developed a national MI strategy and establishedthe Mauritius MI Council and the Mauritius MI Academy to support the de-velopment and implementation of MI technology in the country. The strategyfocuses on how MI can support the ocean economy, which comprises over 10%of Mauritius’ GDP. For example, it suggests investment into a maritime Inter-net of Things (IoT). The strategy also established an MI Council that advisesthe government on supporting Mauritius’ MI ecosystem. Both the MI Strategyand the Mauritius 2030 Strategic Plan prioritize developing local talent, suchas through making programming a required university course.6.10 MozambiqueMozambique2020-2023 Concrete ActionsResearch ✓ waterSMB ✓Table 46: MI in Mozambique1156.10.1 Research[225] examines the quality of OpenStreetMap in Mozambique using unsuper-vised machine learning. Anyone can contribute geographic information to Open-StreetMap (OSM), regardless of their level of experience or skills, which hasraised concerns about quality. When reference data is not available to assessthe quality of OSM data, intrinsic methods that assess the data and its meta-data can be used. In this study, we applied unsupervised machine learning foranalysing OSM history data to get a better understanding of who contributedwhen and how in Mozambique. Even though no absolute statements can bemade about the quality of the data, the results provide valuable insight intothe quality. Most of the data in Mozambique (93%) was contributed by a smallgroup of active contributors (25%). However, these were less active than theOSM Foundation’s definition of active contributorship and the HumanitarianOpenStreetMap Team (HOT) definition for intermediate mappers. Comparedto other contributor classifications, our results revealed a new class: contrib-utors who were new in the area and most likely attracted by HOT mappingevents during disaster relief operations in Mozambique in 2019. More studies indifferent parts of the world would establish whether the patterns observed hereare typical for developing countries. Intrinsic methods cannot replace groundtruthing or extrinsic methods, but provide alternative ways for gaining insightabout quality, and they can also be used to inform efforts to further improvethe quality. We provide suggestions for how contributor-focused intrinsic qualityassessments could be further refinedAccurate land use and land cover (LULC) mapping is essential for scientificand decision-making purposes. [226] maps LULC classes in the northern regionof Mozambique between 2011 and 2020 based on Landsat time series processedby the Random Forest classifier in the Google Earth Engine platform. Thefeature selection method was used to reduce redundant data. The final mapscomprised five LULC classes (non-vegetated areas, built-up areas, croplands,open evergreen and deciduous forests, and dense vegetation) with an overallaccuracy ranging from 80.5% to 88.7%. LULC change detection between 2011and 2020 revealed that non-vegetated areas had increased by 0.7%, built-upby 2.0%, and dense vegetation by 1.3%. On the other hand, open evergreenand deciduous forests had decreased by 4.1% and croplands by 0.01%. The ap-proach used in this paper improves the current systematic mapping approach inMozambique by minimizing the methodological gaps and reducing the temporalamplitude, thus supporting regional territorial development policies.[227] provides an overview of the most important practices in the field of MIused in educational contexts, with a focus on the main platforms used for teach-ing (LMS) to support the development of a research work at Eduardo MondlaneUniversity in Mozambique. To that end, definitions and descriptions of relevantterms, a brief historical overview of MI in education and an overview of thecommon goals and practices of using computational methods in educationalcontexts are provided. The state of the art regarding the adaptation and use ofMI is presented and we discuss the potential benefits and the open challenges.116[228] introduces food security and advanced imaging radiometer datasets andML models. As part of the chapter, satellite radiometer, dairy, food security andsatellite data, global vegetation - Cropland and Vegetation Index, and the Nor-malized Difference Vegetation Index are also covered. Their work also coversMozambique cashew nuts market, agriculture, and industrialization. It con-cludes with two machine learning models that specifically look at Mozambiquecashew nuts production model and Mozambique cashew nuts and NormalizedDifference Vegetation Index model.[229] maps smallholder irrigated agriculture in sub-Saharan Africa using re-mote sensing techniques is challenging due to its small and scattered areas andheterogenous cropping practices. A study was conducted to examine the impactof sample size and composition on the accuracy of classifying irrigated agricul-ture in Mozambique’s Manica and Gaza provinces using three algorithms: ran-dom forest (RF), support vector machine (SVM), and Machine neural network(ANN). Four scenarios were considered, and the results showed that smallerdatasets can achieve high and sufficient accuracies, regardless of their compo-sition. However, the user and producer accuracies of irrigated agriculture doincrease when the algorithms are trained with larger datasets. The study alsofound that the composition of the training data is important, with too few or toomany samples of the “irrigated agriculture” class decreasing overall accuracy.The algorithms’ robustness depends on the training data’s composition, withRF and SVM showing less decrease and spread in accuracies than ANN. Thestudy concludes that the training data size and composition are more importantfor classification than the algorithms used. RF and SVM are more suitable forthe task as they are more robust or less sensitive to outliers than the ANN.Overall, the study provides valuable insights into mapping smallholder irrigatedagriculture in sub-Saharan Africa using remote sensing techniques.[230] uses deep learning and Twitter for mapping in Mozambique. Accu-rate and detailed geographical information digitizing human activity patternsplays an essential role in response to natural disasters. Volunteered geograph-ical information, in particular OpenStreetMap (OSM), shows great potentialin providing the knowledge of human settlements to support humanitarian aid,while the availability and quality of OSM remains a major concern. The ma-jority of existing works in assessing OSM data quality focus on either extrinsicor intrinsic analysis, which is insufficient to fulfill the humanitarian mappingscenario to a certain degree. This paper aims to explore OSM missing built-upareas from an integrative perspective of social sensing and remote sensing. First,applying hierarchical DBSCAN clustering algorithm, the clusters of geo-taggedtweets are generated as proxies of human active regions. Then a deep learningbased model fine-tuned on existing OSM data is proposed to further map themissing built-up areas. Hit by Cyclone IdMI and Kenneth in 2019, the Republicof Mozambique is selected as the study area to evaluate the proposed methodat a national scale. As a result, 13 OSM missing built-up areas are identifiedand mapped with an over 90% overall accuracy, being competitive comparedto state-of-the-art products, which confirms the effectiveness of the proposedmethod.117[231] compares seven machine learning algorithms. Logistic regression (LR)is the most common prediction model in medicine. In recent years, supervisedmachine learning (ML) methods have gained popularity. However, there aremany concerns about ML utility for small sample sizes. In this study, we aim tocompare the performance of 7 algorithms in the prediction of 1-year mortalityand clinical progression to AIDS in a small cohort of infants living with HIVfrom South Africa and Mozambique. The data set (n = 100) was randomly splitinto 70% training and 30% validation set. Seven algorithms (LR, Random For-est (RF), Support Vector Machine (SVM), K-Nearest Neighbor (KNN), NaiveBayes (NB), Machine Neural Network , and Elastic Net) were compared. Thevariables included as predictors were the same across the models including so-ciodemographic, virologic, immunologic, and maternal status features. For eachof the models, a parameter tuning was performed to select the best-performinghyperparameters using 5 times repeated 10-fold cross-validation. A confusion-matrix was built to assess their accuracy, sensitivity, and specificity. RF rankedas the best algorithm in terms of accuracy (82,8%), sensitivity (78%), and AUC(0,73). Regarding specificity and sensitivity, RF showed better performancethan the other algorithms in the external validation and the highest AUC. LRshowed lower performance compared with RF, SVM, or KNN. The outcome ofchildren living with perinatally acquired HIV can be predicted with consider-able accuracy using ML algorithms. Better models would benefit less specializedstaff in limited resources countries to improve prompt referral in case of high-riskclinical progression.[232] focuses on machine learning aspects of Bantu language Emakhuwa ofMozambique. Major advancement in the performance of machine translationmodels has been made possible in part thanks to the availability of large-scaleparallel corpora. But for most languages in the world, the existence of such cor-pora is rare. Emakhuwa, a language spoken in Mozambique, is like most Africanlanguages low-resource in NLP terms. It lacks both computational and linguisticresources. In this paper we describe the creation of the Emakhuwa-Portugueseparallel corpus, which is a collection of texts from the Jehovah’s Witness web-site and a variety of other sources including the African Story Book website,the Universal Declaration of Human Rights and Mozambican legal documents.The dataset contains 47,415 sentence pairs, amounting to 699,976 word tokensof Emakhuwa and 877,595 word tokens in Portuguese. After normalization pro-cesses which remain to be completed, the corpus will be made freely availablefor research use.6.11 Rwanda6.11.1 ResearchThe work in [306] examines rainfall-induced landslide prediction in NgororeroDistrict, Rwanda. The study described in [307] focuses on predicting out-of-pocket health expenditures in Rwanda using machine learning techniques. [308]presents a case study on modeling and mapping soil nutrient depletion in the118Rwanda 2020-2023 Concrete ActionsResearch ✓ waterSMB ✓ AQUA SAFI, TabiriAnalyticsInformalEconomy✓Government✓ Africa’s Centre of Ex-cellence in Artificial In-telligenceTable 47: MI in Rwandahumid highlands of East Africa, specifically in Rwanda, using ensemble ma-chine learning. The paper in [309] employs a machine learning approach topredict the demand for essential medicines in Rwanda based on consumptiondata. The research outlined in [310] utilizes machine learning techniques forpredicting stunting among under-5 children in Rwanda. [311] explores the earlydetection of students at risk of poor performance in Rwanda’s higher educationusing machine learning techniques. In [312], the authors compare different ma-chine learning classifiers to predict hospital readmission of heart failure patientsin Rwanda. [313] focuses on predicting landslide susceptibility and risks in theupper Nyabarongo catchment of Rwanda using GIS-based machine learning sim-ulations. The study detailed in [314] uses machine learning and remote sensingto value property in Rwanda. [315] introduces a machine learning-based triagetool for children with acute infection in a low-resource setting in Rwanda. [316]employs machine learning and the Internet of Things for malaria outbreak pre-diction in Rwanda. The research in [317] predicts crop yields for Irish potatoand maize in Rwanda using machine learning models. [318] compares super-vised machine learning algorithms for road traffic crash prediction models inRwanda. [319] proposes a data-driven predictive machine learning model forefficiently storing temperature-sensitive medical products, such as vaccines, inRwandan pharmacies. [320] applies deep learning techniques to estimate green-house gases emissions from agricultural activities in Rwanda. The work in [321]focuses on creating farmers’ awareness of fall armyworms pest detection at anearly stage in Rwanda using deep learning. [322] utilizes UAV-based mapping tosupport decision-making for banana cultivation in Rwanda. [323] introduces aconvolutional neural network for checkbox detection on Rwandan perioperativeflowsheets. [324] provides ground truths to support remote-sensing inference ofirrigation benefits and effects in Rwanda. [325] compares tree-based models andlogistic regression classifiers for predicting business success in Rwanda. [326] in-troduces an ensemble mode decomposition combined with SVR-RF model forpredicting groundwater levels in Eastern Rwandan aquifers. [327] explores IoTand ML-based precision agriculture, focusing on Rwanda’s coffee industry. Theresearch in [328] leverages convolutional neural networks and satellite images to119map informal settlements in urban settings of the city of Kigali, Rwanda.6.11.2 Small BusinessesAqua Safi assists fish farmers in improving fish productivity by providing asystem to check water quality, manage fish feeding, and adopt best practices forfisheries yield. Tabiri Analytics is a cybersecurity company building the firstaffordable, comprehensive, and automated cybersecurity-as-a-service solutionfor enterprises in underserved markets, utilizing machine learning.6.11.3 GovernmentIn 2023, the Cabinet of the Republic of Rwanda has approved the NationalArtificial Intelligence Policy, which The Future Society supported in draftingfrom 2020 to 2021. The Office of the Prime Minister announced the Policy’sapproval in a Cabinet resolution communiqué in April 2023. The National Ar-tificial Intelligence Policy for the Republic of Rwanda serves as a roadmap toenable Rwanda to harness the benefits of MI and mitigate its risks. Buildingon the mission of the Vision 2050, Smart Rwanda Master Plan and other keynational plans and policies, it equips Rwanda to harness MI for sustainable andinclusive growth. By mobilizing local, regional, and international stakeholders,it positions Rwanda to become a leading African Innovation Hub and Africa’sCentre of Excellence in Artificial Intelligence. The National AI Policy has beendeveloped by MINICT and RURA, with support by GIZ FAIR Forward, theCentre for the 4th Industrial Revolution Rwanda (C4IR) and The Future So-ciety. The National AI Policy, which promotes and fosters Rwanda’s inclusiveand sustainable socio-economic transformation, is oriented around the follow-ing vision and mission statements. Bringing together key national stakeholders,including line ministries, regulation authorities, academia, private sector, start-ups, CSOs and development partners, the two-day workshop was held at LemigoHotel in Kigali, Rwanda from 26 to 27 September 2023. It was organized bythe Rwandan National Commission for UNESCO in collaboration with the UN-ESCO Regional Office for Eastern Africa, the Ministry of ICT & Innovation(MINICT), the German National Commission for UNESCO, Rwanda Devel-opment Board, University of Rwanda, the National Council for Science andTechnology, and the Rwanda Information Society Authority. The UNESCORecommendation on the Ethics of AI, adopted in November 2023 by all 193UNESCO member states, formed the basis of the national workshop, whichfocused on the implementation of the Recommendation in Rwanda.6.12 Seychelles6.12.1 ResearchThe research in [329] investigates the geomorphological drivers influencing deeperreef habitats around Seychelles, providing insights into the underwater land-scape. [330] presents a study mapping the national seagrass extent in Seychelles120Seychelles 2020-2023 Concrete ActionsResearch ✓ D. sechelliaSMB ✓InformalEconomy✓Government✓ National AI StrategyTable 48: MI in Seychellesusing PlanetScope NICFI data, contributing to the understanding of coastalecosystems in the region. In [331], supervised machine learning is employed toreveal introgressed loci in the genomes of Drosophila simulans and D. sechellia,shedding light on genetic interactions between these species. [332] introducesBayesian models for multiple outcomes, with an application to the SeychellesChild Development Study, offering a statistical framework for analyzing diversefactors influencing child development in Seychelles.6.12.2 GovernmentIn 2019, the government of Seychelles launched the National AI Strategy, whichaims to promote the development and adoption of MI in Seychelles while en-suring that the benefits are shared equitably. The strategy focuses on four keyareas: education and skills development, research and innovation, regulatoryframework, and ethical considerations.6.13 SomaliaSomalia 2020-2023 Concrete ActionsResearch ✓SMB ✓InformalEconomy✓Government✓Table 49: MI in Somalia6.13.1 ResearchJetson is an experimental project launched by UNHCR’s Innovation Service in2017 to better understand how data can be used to predict movements of peoplein Sub-Saharan Africa, particularly in the Horn of Africa. The project combinesdata science, statistical processes, design-thinking techniques, and qualitativeresearch methods. It actively seeks new data sources, new narratives, and new121collaborations in order to keep iterating, and improving. Jetson initially fo-cused on understanding the catalysts that cause people to flee their homes inSomalia. Extensive field research resulted in the definition of ten key variablesof forced displacement, such as commodity market prices, rainfall, and violentconflicts. Supported by machine learning, these variables inform an index thatallows for short-term predictions of expected migration flows out of Somalia. Tofulfil its mission, Jetson works in collaboration with partners such as the WorldMeteorological Organization, the Met Office in the UK, academia, and otherUN institutions such as UN Global Pulse. Overall, Project Jetson demonstratesan innovative use of machine learning in the context of forced migration move-ments: It runs short-term predictions more efficiently, at a higher frequency,and at lower costs than traditional calculations. Potentially, the project canbe replicated to other contexts that currently are regions of frequent forcedmigration out-flows.Understanding and analysing sentiment in user-generated content has be-come crucial with the increasing use of social media and online platforms. How-ever, sentiment analysis in less-resourced languages like Somali poses uniquechallenges. The work in [333] test presents the performance of three ML al-gorithms (DTC, RFC, XGB) and two DL models (CNN, LSTM) in accuratelyclassifying sentiment in Somali text. The CC100-Somali dataset, comprising78M monolingual Somali texts from the Common crawl snapshots, is utilizedfor training and evaluation. The study employed rigorous evaluation techniques,including train-test splits and cross-validation, to assess classification accuracyand performance metrics. The results demonstrated that DTC achieved thehighest accuracy among ML algorithms, 87.94%, while LSTM achieved thehighest accuracy among DL models, 88.58%. This study’s findings contribute tosentiment analysis in less-resourced languages, specifically Somali, and providevaluable insights into the performance of ML and DL techniques. Moreover,the study highlights the potential of leveraging both ML and DL approachesto analyze sentiment in Somali text effectively. The results and evaluationmetrics benchmark future research in sentiment analysis for Somali and otherlow-resource languages.6.14 South SudanSouth Su-dan2020-2023 Concrete ActionsResearch ✓SMB ✓InformalEconomy✓Government✓Table 50: MI in South Sudan1226.14.1 Research[334] introduces and applies a set of machine intelligence techniques to analyzemulti-dimensional fragility-related data. Our analysis of the fragility data col-lected by the OECD for its States of Fragility index showed that the use of suchtechniques could provide further insights into the non-linear relationships anddiverse drivers of state fragility, highlighting the importance of a nuanced andcontext-specific approach to understanding and addressing this multi-aspect is-sue. They applied the methodology to South Sudan, one of the most fragilecountries in the world to analyze the dynamics behind the different aspects offragility over time. The results could be used to improve the Fund’s countryengagement strategy (CES) and efforts in the country.6.15 TanzaniaTanzania 2020-2023 Concrete ActionsResearch ✓ waterSMB ✓InformalEconomy✓Government✓Table 51: MI in Tanzania6.15.1 ResearchThe study in [335] aims to explore the current status, challenges, and opportu-nities for MI application in the health system in Tanzania. A scoping reviewwas conducted using the Preferred Reporting Items for Systematic Review andMeta-Analysis Extensions for Scoping Review (PRISMA-ScR). They searcheddifferent electronic databases such as PubMed, Embase, African Journal Online,and Google Scholar. Eighteen (18) studies met the inclusion criteria out of 2,017studies from different electronic databases and known MI-related project web-sites. Amongst MI-driven solutions, the studies mostly used machine learningand deep learning for various purposes, including prediction and diagnosis ofdiseases and vaccine stock optimisation. The most commonly used algorithmswere conventional machine learning, including Random Forest and Neural net-work, Naive Bayes K-Nearest Neighbour and Logistic regression. This reviewshows that MI-based innovations may have a role in improving health servicedelivery, including early outbreak prediction and detection, disease diagnosisand treatment, and efficient management of healthcare resources in Tanzania.Their results indicate the need for developing national MI policies and regula-tory frameworks for adopting responsible and ethical AI solutions in the health123sector in accordance with the World Health Organisation guidance on ethicsand governance of MI for health.In Dar es Salaam and many other cities, poorer neighborhoods tend to haveless vegetation and tree canopy around them. Take Namanga, a densely packedinformal settlement in eastern Dar es Salaam. Despite abutting some of thecity’s wealthiest and greenest neighborhoods, residents of Namanga must en-dure weather extremes with barely any green cover to provide shade or absorbrun-off from sudden downpours. Understanding which districts have ample orinadequate tree canopy cover is challenging but it is becoming more feasibledue to increasingly detailed satellite imagery. Machine learning algorithms area key technology for interpreting such imagery. Still, algorithms to detect fea-tures such as tree canopy are only as good as the data they are built on. Whentheir team applied an off-the-shelf tree detection algorithm (developed usinga tree canopy dataset from California) to satellite imagery of Dar es Salaam,the results were unsatisfactory [336]. Data labeling is the process of addingmeaningful information to raw data so that computers can learn to recognizepatterns in it, for example, annotating recordings of human speech with thewords they contain or identifying objects in photographs. To produce detailedtree-cover maps of Dar es Salaam and Freetown, the capital of Sierra Leone,the Resilience Academy students began by developing their own large datasetof labeled satellite imagery. Using an open-source labeling tool developed byAzavea, the students loaded high-resolution satellite imagery, divided it intogrid cells, and drew accurate boundaries around the tree canopy. By labelingjust 1% of the city in this way, the resulting dataset enabled a machine learningmodel to learn how to recognize its trees, distinguishing tree canopy from grass,buildings and other features, even in shady conditions.6.16 UgandaUganda 2020-2023 Concrete ActionsResearch ✓ speech datasetsSMB ✓ Chil Ai Lab, GlobalAuto Systems, Weke-bereInformalEconomy✓Government✓ Artificial Intelligenceand Data Science LabTable 52: MI in Uganda1246.16.1 ResearchThe project [337] aims deliver open, accessible, and high-quality text and speechdatasets for low-resource East African languages from Uganda, Tanzania, andKenya. Taking advantage of the advances in NLP and voice technology requiresa large corpora of high quality text and speech datasets. This project will aimto provide this data for these languages: Luganda, Runyankore-Rukiga, Acholi,Swahili, and Lumasaaba. The speech data for Luganda and Swahilli will begeared towards training a speech-to-text engine for an SDG relevant use-caseand general-purpose ASR models that could be used in tasks such as drivingaids for people with disabilities and development of MI tutors to support earlyeducation. Monolingual and parallel text corpora will be used in several NLPapplications that need NLP models, including natural language classification,topic classification, sentiment analysis, spell checking and correction, and ma-chine translation.6.16.2 Small BusinessesChil AI utilizes Machine Learning and Artificial Intelligence to offer Telehealthservices, Electronic medical records, E-consultation, automated laboratory re-sults interpretation, E-referral, and E-pharmacy services to African women.Global Auto Systems aims to revolutionize the healthcare system in Ugandaby using AI and Cloud Computing technologies to improve patient outcomeswhile reducing the total cost of care. Wekebere is a health social enterprisestriving to engineer innovative healthcare solutions that give expectant mothersin low-resource settings the healthy lives they deserve.6.16.3 GovernmentThe Artificial Intelligence and Data Science lab in Uganda specializes in the ap-plication of artificial intelligence and data science - including, for example, meth-ods from computer vision, natural language processing and predictive analytics-to problems in the developing world. Applications include Natural languageprocessing for under-resourced languages, automated diagnosis of both cropand human diseases, auction design for mobile commodity markets, analysis oftraffic patterns in African cities, and of telecoms and remote sensing data foranticipating the spread of infectious diseases6.17 Zambia6.17.1 ResearchStunting is a global public health issue. We sought to train and evaluate machinelearning (ML) classification algorithms on the Zambia Demographic Health Sur-vey (ZDHS) dataset to predict stunting among children under the age of five inZambia. The authors of [287] applied Logistic regression , Random Forest, SV125Zambia 2020-2023 Concrete ActionsResearch ✓ MiningSMB ✓InformalEconomy✓Government✓Table 53: MI in Zambiaclassification, XG Boost (XgB) and Naive Bayes algorithms to predict the prob-ability of stunting among children under five years of age, on the 2018 ZDHSdataset. We calibrated predicted probabilities and plotted the calibration curvesto compare model performance. We computed accuracy, recall, precision and F1for each machine learning algorithm. About 2327 (34.2%) children were stunted.Thirteen of fifty-eight features were selected for inclusion in the model using ran-dom forest. Calibrating the predicted probabilities improved the performanceof machine learning algorithms when evaluated using calibration curves. RFwas the most accurate algorithm, with an accuracy score of 79% in the testingand 61.6% in the training data while Naive Bayesian was the worst performingalgorithm for predicting stunting among children under five in Zambia using the2018 ZDHS dataset. ML models aids quick diagnosis of stunting and the timelydevelopment of interventions aimed at preventing stunting.[288] examines audits in Zambia. Independent verification is a critical com-ponent of performance-based financing (PBF) in health care, in which facilitiesare offered incentives to increase the volume of specific services but the sameincentives may lead them to over-report. We examine alternative strategies fortargeted sampling of health clinics for independent verification. Specifically, weempirically compare several methods of random sampling and predictive mod-eling on data from a Zambian PBF pilot that contains reported and verifiedperformance for quantity indicators of 140 clinics. Our results indicate thatmachine learning methods, particularly Random Forest, outperform other ap-proaches and can increase the cost-effectiveness of verification activities.Air quality monitoring in Zambian mining towns is an important issue dueto the high levels of pollution caused by mining activities. Zambia is a countryrich in minerals and mining is a significant contributor to its economy. How-ever, mining activities have also led to increased levels of air pollution in miningtowns, affecting the health of local communities. According to the Ministry ofMines, the major sources of air pollution in the Copperbelt are smelters, min-ing, and quarrying among others. Additionally, the Ministry of Mines reportsthat major pollutants include sulfur dioxide (SO2), oxides of nitrogen (NOx),particulate matter, carbon monoxide (CO), dust, Carbon dioxide, etc. Thereare several government agencies engaged in management that can help withthese environmental issues, including the Zambia Environmental ManagementAgency (ZEMA). A research was investigated in [289] by using a thorough review126of the literature, furthermore, a qualitative study was conducted at ZEMA theprimary institution for environmental monitoring, and specifically, interviewswere conducted. This was done in order to gain an in-depth overview of thecurrent state of the art for environmental pollutant monitoring in affected min-ing towns. According to the findings presented here, the country has not madeenough investments in environmental monitoring technologies and instead relieson funded projects that render the agency responsible for preventing and con-trolling ambient pollution inoperable after the projects are completed, despitethe fact that there are plenty of mineral resources available and more are still tobe discovered. The research suggested new techniques for comparing ambientair pollutant levels to national guideline limits based on the limitations of itsresults. This study uses data from an ongoing obstetrical cohort in Lusaka,Zambia that uses early pregnancy ultrasound to estimate GA. Our intent wasto identify the best set of parameters commonly available at delivery to cor-rectly categorize births as either preterm (37 weeks) or term, compared to GAassigned by early ultrasound as the gold standard. Trained midwives conducteda newborn assessment (72 hours) and collected maternal and neonatal dataat the time of delivery or shortly thereafter. New Ballard Score (NBS), lastmenstrual period (LMP), and birth weight were used individually to assign GAat delivery and categorize each birth as either preterm or term. Additionally,machine learning techniques incorporated combinations of these measures withseveral maternal and newborn characteristics associated with prematurity andSGA to develop GA at delivery and preterm birth prediction models. The dis-tribution and accuracy of all models were compared to early ultrasound dating.Within our live-born cohort to date (n = 862), the median GA at delivery byearly ultrasound was 39.4 weeks. Among assessed newborns with complete dataincluded in this analysis (n = 468), the median GA by ultrasound was 39.6weeks. Using machine learning, we identified a combination of six accessibleparameters (LMP, birth weight, twin delivery, maternal height, hypertension inlabor, and HIV serostatus) that can be used by machine learning to outperformcurrent GA prediction methods. For preterm birth prediction, this combinationof covariates correctly classified 94% of newborns and achieved an area underthe curve (AUC) of 0.9796. We identified a parsimonious list of variables thatcan be used by machine learning approaches to improve accuracy of pretermnewborn identification. Our best-performing model included LMP, birth weight,twin delivery, HIV serostatus, and maternal factors associated with SGA. Thesevariables are all easily collected at delivery, reducing the skill and time requiredby the frontline health worker to assess GA.The work in [290] presents a method to identify poor households in data-scarce countries by leveraging information contained in nationally representativehousehold surveys. It employs standard statistical learning techniques - cross-validation and parameter regularization - which together reduce the extent towhich the model is over-fitted to match the idiosyncracies of observed surveydata. The automated framework satisfies three important constraints of thisdevelopment setting: i) The prediction model uses at most ten questions, whichlimits the costs of data collection; ii) No computation beyond simple arithmetic127is needed to calculate the probability that a given household is poor, immedi-ately after data on the ten indicators is collected; and iii) One specification of themodel (i.e. one scorecard) is used to predict poverty throughout a country thatmay be characterized by significant sub-national differences. Using survey datafrom Zambia, the model’s out-of-sample predictions distinguish poor householdsfrom non-poor households using information contained in ten questions.Assessing tax gaps - the difference between the potential and actual taxesraised - plays a vital role in achieving positive domestic revenue objectivesthrough improved and reformed taxation. This is particularly pertinent forgrowth outcomes in developing countries. This study in [291] uses a bottom-upapproach based on micro-level audit information to estimate the extent of taxmisreporting in Zambia. Our methods predict the extent of tax evasion using aregression and a machine learning algorithm based on a sample of audited firms,after which we estimate tax gaps using a standard approach. We estimate totaltax gaps as 56 per cent and 47 per cent for the two approaches, respectively.These gaps are mainly driven by corporate taxes. Applying our gap to key in-dustries shows that the extractives sector in Zambia records the highest gaps interms of CIT and one of the lowest gaps in terms of VAT.The World Health Organization recommends chest radiography to facilitatetuberculosis (TB) screening. However, chest radiograph interpretation exper-tise remains limited in many regions. [292] developed a deep learning system(DLS) to detect active pulmonary TB on chest radiographs and compare itsperformance to that of radiologists. A DLS was trained and tested using retro-spective chest radiographs (acquired between 1996 and 2020) from 10 countries.To improve generalization, large-scale chest radiograph pretraining, attentionpooling, and semi-supervised learning (“noisy-student”) were incorporated. TheDLS was evaluated in a four-country test set (China, India, the United States,and Zambia) and in a mining population in South Africa, with positive TB con-firmed with microbiological tests or nucleic acid amplification testing (NAAT).The performance of the DLS was compared with that of 14 radiologists. Theauthors studied the efficacy of the DLS compared with that of nine radiologistsusing the Obuchowski-Rockette-Hillis procedure. A deep learning method wasfound to be noninferior to radiologists for the determination of active tubercu-losis on digital chest radiographs.Deep Learning System is used in [293] to Screen for Diabetic Retinopathyin an Underprivileged African Population with Diabetes. Diabetes exerts anemerging burden in Zambia. Related complications such as diabetic retinopa-thy (DR) are expected to increase dramatically in prevalence. Challenged byshortage of ophthalmic services and poor accessibility to DR screening, the ap-plication of machine intelligence ( using deep learning may be an alternativesolution. This study aims to evaluate the real-world clinical effectiveness ofa DL system in screening for DR and vision-threatening DR (VTDR) in theZambian population with diabetes. A total of 4513 images from 3101 eyes of1578 Zambians with diabetes were prospectively recruited for this study. Two-field color 45-degree retinal fundus photographs were captured for each eye andgraded according to International Classification of Diabetic Retinopathy Sever-128ity scale. Referable DR was defined as moderate non-proliferative DR (NPDR)or worse, diabetic macular edema and ungradable images; VTDR was designatedas severe NPDR and proliferative DR. With reference to the retinal specialists’grading, we calculated the area under the receiver operating curve (AUC), sensi-tivity and specificity for referable DR, and the detection rate of VTDR, using anEnsemble convolutional neural network. The developed DL system shows clin-ically acceptable performance in detection of referable DR and VTDR for theZambian population. This demonstrates the potential application to adopt suchsophisticated cutting-edge MI technology for the underprivileged population.6.18 ZimbabweZimbabwe 2020-2023 Concrete ActionsResearch ✓ NeotreeSMB ✓InformalEconomy✓Government✓Table 54: MI in Zimbabwe6.18.1 ResearchEfficient and effective healthcare systems utilize the available data at every levelto provide evidence-based care and improve procedures and practice in order tomeet the three goals of healthcare institutions - access, quality and efficiency.Regardless of the changing child health needs and often failure by traditionalhealthcare models to cope, most of the public health institutions in Sub-SaharanAfrica are not providing data-driven pediatrics to meet the ever-changing childhealth needs. There is a lack of utilization of the data collected by the routinehealth information systems to provide evidence-based and personalized pedi-atrics. The study in [294] employed an exploratory research design to explorethe opportunities, and potential challenges of data-driven pediatrics based onthe lessons learnt from the introduction of the electronic health record andNeotree (a digital health system deployed in Zimbabwe and Malawi to helphealth workers manage neonates’ health) in Makonde District. Twenty publichealth workers participated in interviews and focus groups and reports fromthe district health information system provided further insights. The study re-vealed that data-driven pediatrics could improve access, efficiency and qualityof pediatric care, regardless of such potential challenges as fear of medico-legalhazards, centralization of decision-making, resistance by healthcare workers,network challenges and computer illiteracy. To increase the chances of success,the following lessons learnt from electronic health record and neotree introduc-tion could help: start small, sensitize communities first, involve line healthcare129workers from the beginning, do not train in a haste and demystify technology’spurpose. The study revealed that there are pediatricians and nurses willing toshift to data-driven pediatrics if the technologies are available.The work in [295] studies vehicle damage. According to the United NationsRoad Safety Performance Review-Zimbabwe report, every 15 minutes, five peo-ple die in road accidents within Zimbabwe, recording the highest number of ac-cidents in the SADC region. The situation has brought more pressure and workin the insurance sector as they are expected to process all the claims accuratelyand timely. Deep learning entails automation, enhancement, analysis, and highaccuracy in areas like speech recognition, object detection, and language trans-lation. In this paper, two modern deep learning algorithms MobileNetV2 andDenseNetV121 were used to develop the vehicle damage classification models.The models were used to detect damaged main features of a car, which are: thedoor, bumper, windscreen, tail lamp, and headlamp. Mobile NetV 2’s53 lay-ers and DenseNet121’s121 layers produced high accuracy rates for identifyingdamaged parts in vehicles. However, DenseNetV2 produced a higher accuracyof 84 & than MobileNetV2, with an accuracy rate of 78%. The models alsoused low computational resources than the traditional algorithms making themapplicable in different insurance companies as they can be easily embedded intoclient’s mobile phones.The work in [296] predicts depression Among Adolescents. Depression beinga behavioural health disorder is a serious health concern in Zimbabwe and allover the world. If depression goes unaddressed, the consequences are detrimen-tal and have an impact on the way one behaves as an individual and at thesocietal level. Despite the number of individuals who could benefit from treat-ment for behavioural health concerns, their difficulties are often unidentified andunaddressed through treatment. Technology carries the unrealized potential toidentify people at risk of behavioural health conditions and to inform preventionand intervention strategies7 On MI adoption metrics and evaluation ofcountries MI strategiesOn MI adoption metricsMany people are referring to the Unified Theory of Acceptance and Use ofTechnology (UTAUT) 1, 2, 3, etc., to explain the adoption of MI, blockchain,CBDC, autonomous vehicles, etc. From a game theory perspective, UTAUTis not universal, and it is not a one-size-fits-all model. To date, there is nosingle method that evaluates the adoption of all technologies in a way that isuniversal and cohesive. UTAUT stands for Unified Theory of Acceptance andUse of Technology. It’s a model that aims to understand how individuals adoptand use new technologies.130What is UTAUT in MI ?In the context of MI, UTAUT can be applied to study factors influencing theacceptance and use of intelligent technologies. The key components of UTAUTin MI typically include:• Performance Expectancy: Users’ belief that using the technology will en-hance their performance.• Effort Expectancy: The perceived ease of use and minimal effort requiredto use the technology.• Social Influence: The impact of social factors and opinions on the user’sdecision to adopt MI.• Facilitating Conditions: The degree to which users believe that the nec-essary resources and support are available for using the technology.• Behavioral Intention: Users’ intention to use MI based on the factorsmentioned.Limitations of UTAUT in MIWhile UTAUT is a valuable framework, it has some limitations in the contextof MI, including:• Lack of Specificity: UTAUT may not capture the unique challenges andnuances associated with different types of MI technologies.• Dynamic Nature: MI evolves rapidly, and UTAUT might struggle to keepup with emerging technologies and user perceptions.• Cultural Variations: The model may not fully account for countries cul-tural differences in the acceptance of MI, as cultural factors play a signif-icant role in technology adoption.• Ethical Concerns: UTAUT primarily focuses on user acceptance but maynot adequately address ethical considerations related to MI, which arecrucial in this domain.• Limited Emotional Factors: The emotional aspect of human-machine in-teraction is not extensively covered, and emotions can significantly influ-ence technology acceptance.• MI researchers often need to complement UTAUT with additional frame-works or consider these limitations when applying it to the rapidly evolvingfield of MI.We provide below specific practical examples in MI where UTAUT fails131• Deep Learning vs. Shallow Learning: UTAUT may not capture the nu-anced differences in user acceptance between deep learning and traditionalshallow learning methods due to the distinctive nature of these technolo-gies.• Explainable MI (XMI): UTAUT might not adequately address the chal-lenges associated with user acceptance of Explainable MI systems, whereusers often prioritize transparency and interpretability over performance.• Humanoid Robots: When it comes to the adoption of humanoid robotspowered by MI, UTAUT may not sufficiently consider the impact of an-thropomorphism and emotional factors on user acceptance.• Biometric Surveillance Systems: UTAUT might fall short in addressingthe concerns and ethical considerations associated with the acceptanceof MI in biometric surveillance systems, where privacy and security playcrucial roles.• Generative Adversarial Networks (GANs): UTAUT may not effectivelycapture the unique challenges in user acceptance of GANs, where thetechnology is used to generate realistic synthetic data, raising concernsabout its ethical implications and potential misuse.• Chatbots: UTAUT may struggle to fully capture the user acceptance dy-namics of chatbots, especially considering factors like natural languageunderstanding, context awareness, and the ability to engage users in mean-ingful conversations.• Conversational Agents in Medicine: When applied to conversational agentsin healthcare, UTAUT might not address the specific concerns related totrust, accuracy, and privacy that are crucial in the adoption of MI-drivenmedical conversational agents.• Audio-to-Audio Bots: UTAUT may not adequately consider the uniquechallenges associated with user acceptance of audio-to-audio bots, wherethe conversion of spoken language to another language or format intro-duces complexities not fully addressed by traditional UTAUT factors.• Augmented Reality in MI: UTAUT may not sufficiently encompass theuser acceptance factors specific to the integration of MI into augmentedreality applications, where factors like user experience in a mixed realityenvironment become significant.• Personalized MI Assistants: UTAUT might not fully account for the in-tricacies of user acceptance in personalized MI assistants that adapt toindividual preferences and behaviors, as the personalization aspect intro-duces additional dimensions beyond the traditional UTAUT framework.132Game Theory outperforms UTAUTA game theory approach for technology adoption can incorporate the followingelements to outperform the standard UTAUT:• Multi-Agent Interactions: Consider the technology adoption as a dynamicgame with multiple interacting agents, such as MI research funding spon-sors, research investors, network of advertisers, MI sponsored ambas-sadors, users, developers, and policymakers. Each agent’s strategy anddecisions impact others, reflecting the complex interactions in the real-world adoption scenario.• Incentive Structures: Introduce incentive structures for users and devel-opers that align with the specific features of the technology. For instance,rewards for developers creating explainable MI or penalties for users en-gaging in malicious use of the technology.• Dynamic Payoffs: Model dynamic payoffs that evolve over time, reflectingthe changing nature of technology and user perceptions. This accommo-dates the rapid advancements in MI and the evolving needs and preferencesof users.• Privacy Considerations: Incorporate privacy as a strategic element in thegame, recognizing its critical role in user decision-making. Agents mayhave different preferences regarding privacy, affecting their strategies andinteractions within the technology adoption game.• Adversarial Scenarios: Account for adversarial scenarios where users ordevelopers may strategically exploit vulnerabilities or engage in unethicalpractices. This helps in understanding and mitigating risks associatedwith technology adoption.By integrating these specificities into a game theory framework, the model canprovide a more comprehensive understanding of the complex dynamics involvedin the adoption of advanced technologies, outperforming the standard UTAUTin capturing the nuances of the interactions between actors and technologies.To date,• UTAUT Not Unified: UTAUT 1,2,3 lack unity in addressing the diversechallenges of technology adoption, suggesting it falls short of providing acohesive framework.• Not Universal: UTAUT is not universal in its applicability, indicating thatits principles may not effectively cover the wide range of technologies anduser contexts.• Not One-Size-Fits-All: UTAUT is not a one-size-fits-all solution, implyingthat it may not sufficiently adapt to the specificities and complexitiesinherent in different technologies and user interactions.133• MFTG Outperforms UTAUT: A simple Mean-Field-Type Game (MFTG)approach surpasses UTAUT in MI, implying that incorporating basic gametheory principles better captures the dynamics of technology adoption.On MI indexes or metrics of countriesMI indexes for countries assess their capabilities, strategies, and developmentin machine intelligence.Some notable MI indexes include:• Global MI Index: Provides a comprehensive assessment of MI readiness,research, development, and application across various countries.• MI Readiness Index: Focuses on a country’s preparedness for MI adoption,considering factors like infrastructure, education, and innovation.• Government MI Readiness Index: Evaluates how well governments arepositioned to adopt and harness the benefits of MI for public services andgovernance.• MI Policy and Practice Index: Assesses the policies and practices relatedto MI governance, ethics, and regulations in different countries.• MI Innovation Index: Measures a country’s innovation in MI, consideringresearch output, startup activity, and investment in MI-related industries.• MI Ethics Index: Evaluates countries based on their commitment to ethi-cal considerations in MI development and deployment, assessing policies,regulations, and adherence to ethical guidelines.• MI Inclusion Index: Measures the inclusivity of MI initiatives within acountry, considering factors such as diversity in MI workforce, accessibilityof MI technologies, and efforts to minimize biases.• MI Education Index: Assesses a country’s educational infrastructure andinitiatives related to MI, including the availability of MI courses, researchprograms, and the integration of MI education at various levels.• MI Industry Adoption Index: Examines the extent to which industrieswithin a country are adopting and integrating MI technologies into theiroperations, providing insights into economic impact and competitiveness.• MI Investment Index: Focuses on the level of financial investment andfunding dedicated to MI research, development, and implementation, re-flecting a country’s commitment to MI advancements.134Limitations of MI indexes in AfricaThe key limitations of MI indexes for countries include:• Data Availability and Quality: The accuracy and reliability of these in-dexes heavily depend on the availability and quality of data. Inconsistentor incomplete data can lead to skewed assessments.• Static Nature: Index rankings are often based on static snapshots, whichmight not capture the dynamic and evolving nature of MI ecosystems.Rapid changes in technology and policies can quickly impact a country’sstanding.• Subjectivity in Metrics: The choice of metrics and their subjective in-terpretation can introduce bias. Different index creators may prioritizecertain factors over others, leading to variations in rankings.• Lack of Standardization: There is a lack of standardized criteria acrossindexes, making it challenging to compare and aggregate results. Differentmethodologies and criteria can lead to discrepancies in rankings.• Inability to Capture Informal MI Activity: Some indexes may struggle toaccount for informal MI activities, grassroots initiatives, or developmentsin non-traditional sectors, potentially underestimating a country’s overallMI landscape.• Ethical Considerations: Certain indexes may not adequately address eth-ical considerations, such as the responsible use of MI or the impact onprivacy and human rights, which are essential aspects of MI development.• Limited Focus: Indexes often focus on specific aspects of MI, such asreadiness or ethics, but may not provide a holistic view. This limitedfocus can overlook important interactions between different componentsof MI ecosystems.• Global Perspective: Some indexes may lack a truly global perspective,focusing more on developed countries and potentially neglecting the con-tributions and challenges faced by emerging economies in the MI space.Current MI indexes are unethical• Lack of Audio2Audio Processing for African Languages: Current MI tech-nologies, including MI indexes, do not adequately address the need forAudio2Audio processing, particularly for African languages. This crucialaspect, which could benefit the local population in Africa, is largely over-looked in existing evaluations of MI readiness and capacity.• Underrepresentation of African Cultures: Existing MI technologies andindexes often fail to incorporate the rich diversity of African cultures. The135absence of cultural nuances and specificities in these evaluations reflectsa significant gap in understanding and catering to the unique contexts ofAfrican societies, limiting the relevance and accuracy of MI rankings forthe continent.• National MI Strategy: Having a report or white paper that compiles aset of recommendations does not necessarily lead to advanced MI country.Having a national MI strategy outlined in a white paper should not carrythe same weight as the actual implementation of MI that is actively usedand beneficial to the local population. Concrete implementation and im-pact on the ground should be given more emphasis than the existence ofa strategic document.• National MI & Robotics Center: The mere establishment of an MI androbotics center by some countries, devoid of actual contributions to lo-cal MI research, training, innovation, and understanding, should not beequated with meaningful progress. Centers that lack substantial engage-ment with local issues, fail to provide MI training for the community, andfocus only on filmographies and metaverse glasses are more about politicalposturing in the MI race than making a tangible impact. Such govern-mental actions should not carry the same weight as actively implementedand utilized MI technologies that address real challenges in Africa.• Ambiguity of Concepts: MI ’readiness’ and ’capacity’ are ambiguous con-cepts that lack standardized definitions, making it challenging to measurethem consistently, especially in Africa.• Inconsistent Proxy Indicators: Current MI indexes rely on inconsistentand divergent proxy indicators to gauge a country’s readiness or capacityfor MI, leading to unreliable comparisons.• Unequal Weighting of Indicators: Some indexes assign equal weight toindicators that represent vastly different national characteristics, creatingan imbalanced evaluation.• Inappropriate Metrics: Metrics such as public investments in MI may notprovide a fair comparison between states, as they do not consider howthese investments are spent or the quality of MI produced in the area, inthe country, within Africa and across the globe.• Lack of Differentiation: Metrics related to the number of ’MI players,’projects, or services often fail to differentiate between large, influentialorganizations and smaller ventures with less impact on MI development.• Unreliable Data Sources: Some key metrics rely on potentially unreliabledata sources, such as a LinkedIn dataset, Twitter posts, or political Ads,that may include distorted claims about MI skills.136• Self-Reporting Bias: Indicators of MI adoption by organizations often relyon self-reporting, and online surveys, introducing potential inaccuracies.• Lack of Transparency: Some rankings lack transparency in their method-ology and raw data, making it difficult to assess the credibility of theirfindings.• Regional Biases: There is evidence of potential regional biases in indexdata, with raw data on journal citations under-representing academic out-put from Africa and the Global South.• Demographic Biases: LinkedIn, Twitter, Wechat, key data sources, mayunder-represent MI penetration in certain regions of Africa due to varia-tions in user rates worldwide.• Under-Representation of Africa and the Global South: African authorsand Global South authors, articles, and journals may be under-representedin citation counts, contributing to potential biases in assessing MI devel-opment.• Limited Data on Methodology: Lack of publicly released detailed method-ology for certain indexes raises concerns about the credibility of their rank-ings.• Overemphasis onWestern Institutions: A disproportionate number of dataand indicators are compiled, sorted, and presented by Western institu-tions, potentially skewing the evaluation towards Western perspectives.• Inadequate Representation of Non-Male Experts: Some indexes’ staffingand advisory bodies lack adequate representation of non-male experts,introducing potential biases in the evaluation process.• Race-Like Approach Encouragement: Some indexes actively encourage arace-like approach to MI policy, potentially promoting competition ratherthan collaborative development.• Overemphasis on Quantity: Metrics focusing on the number of MI projectsor services may prioritize quantity over quality, neglecting the impact andinnovation of each project. In Africa, the quality should be more usefulto the local population.• Limited Contextual Understanding: Indexes may lack a nuanced under-standing of the local context, leading to misinterpretations and misrepre-sentations of a country’s MI development.• Technological Complexity Oversimplification: The complexity of MI de-velopment cannot be adequately captured by a single ranking, oversimpli-fying the intricate factors involved.137• Insufficient Adaptability: MI indexes may struggle to adapt to the evolvinglandscape of MI development, rendering them less effective in providingaccurate and relevant assessments over time.For all these reasons, Current MI indexes, MI readiness, and MI rankings inAfrica are currently unethical methodologies. We should prioritize making themethical and propose alternative methodologies that account for the specificitiesof each country.8 Multi-Scale MI Ethics in AfricaHere, Ethics refers to the study of moral principles, values, and conduct thatguide individuals, groups, or societies in distinguishing between right and wrongactions. It provides a framework for evaluating the morality of decisions andbehaviors, shaping standards of behavior based on concepts such as fairness, jus-tice, honesty, and integrity. Ethics encompasses rich notions that guide moraldecision-making and behavior. At its core are moral principles such as honesty,integrity, and justice, which underpin ethical frameworks. Values play a crucialrole, representing core beliefs that influence priorities and choices. Rights, bothindividual and collective, emphasize entitlements and freedoms. Duties andresponsibilities arise from societal expectations and professional roles. Virtues,including compassion and empathy, contribute to positive character traits. Con-sequences and outcomes provide a lens for evaluating actions, aligning with eth-ical theories like utilitarianism. Fairness and justice ensure equitable treatment,while integrity underscores consistency and honesty. Autonomy recognizes in-dividuals’ right to independent decisions, and respect acknowledges the intrin-sic value of each person. Other notions include accountability, beneficence,nonmaleficence, empathy, sympathy, cultural competence, beneficence, confi-dentiality, informed consent, social responsibility, empathy, altruism, pluralism,paternalism, reciprocity, and sustainability, collectively forming the intricatefabric of ethical considerations.The current practices of MI ethics in Africa are not ethicalPaternalism is a concept in ethics and governance that involves a person or au-thority making decisions or taking actions on behalf of others with the belief thatit is for their best interest, often without their explicit consent. This approachis rooted in a sense of protection or guidance, assuming that the paternalisticparty knows what is best for the individuals or groups involved. While pater-nalism may be well-intentioned, it can raise ethical concerns as it may infringeon individual autonomy and decision-making. Critics argue that individualsshould have the freedom to make their own choices, even if those choices in-volve potential risks or mistakes. Paternalism is often debated in contexts suchas engineering and healthcare, where decisions about medical treatments andengineering threshold designs may be made by healthcare professionals and en-gineers for the patient’s and users presumed well-being.138The assessment of what is ethical can vary across cultures, philosophical per-spectives, and individual beliefs, but it generally involves the adherence to prin-ciples that promote positive and virtuous conduct. MI ethics is a multifaceteddomain encompassing numerous notions to guide the responsible developmentand deployment of MI. Fundamental principles include fairness, ensuring unbi-ased outcomes and mitigating algorithmic biases. Transparency is vital, empha-sizing the need for clear explanations of MI decision-making processes. Account-ability holds developers and organizations responsible for the ethical impact ofMI systems. Privacy concerns address the responsible handling of personaldata, while security focuses on preventing misuse and safeguarding against ma-licious activities. Bias mitigation is crucial for identifying and rectifying biasesin MI algorithms, promoting equitable results. Explainability ensures that MIdecisions are understandable to diverse stakeholders. Human-centered designprioritizes the well-being and interests of humans in MI development. Inclusiv-ity advocates for considering diverse perspectives and benefiting a broad rangeof communities. Environmental impact considerations assess and minimize theecological footprint of MI systems. Additional notions include autonomy, con-sent, data governance, algorithmic transparency, algorithmic accountability, so-cial responsibility, cultural sensitivity, global collaboration, ethical governance,human augmentation ethics, collaborative intelligence, digital divide reduction,responsible MI education, ethical data sourcing, dual-use technology awareness,ethical advertising, participatory design, ethical MI research, safety conscious-ness, MI for social good, interdisciplinary collaboration, long-term impact as-sessment, value alignment, and continuous stakeholder engagement, collectivelyshaping the ethical landscape of MI.Deploying MI technologies in certain countries in Africa without the consentof the population, authorities, and in a language not widely spoken is generallyconsidered unethical. Ethical MI deployment involves respecting the autonomyand rights of individuals and communities, obtaining informed consent, and en-suring that technologies are culturally and linguistically appropriate. Failureto seek consent and address linguistic and cultural considerations may lead toissues such as lack of user understanding, potential biases, and violations ofprivacy and human rights. Ethical MI practices prioritize transparency, inclu-sivity, and the well-being of the affected communities, and any deployment thatneglects these considerations raises significant ethical concerns.The current MI is not inclusive in AfricaIn its current form, MI is not inclusive as it excludes a significant portion ofAfrican language users who predominantly engage with audio content. Con-sequently, there is a lack of MI ethics in most African countries. Mere copy-pasting of ethics guidelines does not inherently render MI ethical in the Africancontext. In practice in most African countries, ethics is multimodal, multi-dimensional, context-dependent, and culturally sensitive. Unlike the globalstereotypes, Africa is multicultural, varying from one region to another andsometimes within the same country. So, what MI ethics can provide a coherent139and culturally respectful guideline?The classical MI ethics guidelines face several limitations, particularly ininformal sectors in Africa. MI ethics guidelines often reflect Western perspec-tives and may not adequately consider the diverse cultural contexts in informalsectors across Africa. Cultural nuances and ethical considerations may varysignificantly, making it challenging to have universal guidelines that addressall cultural intricacies. Informal sectors in Africa often experience a significantdigital divide, with limited access to technology and digital literacy. MI ethicsguidelines may assume a certain level of technological infrastructure and aware-ness, making their application challenging in contexts where such resources arescarce. Many MI ethics guidelines are presented in languages that may not beaccessible or easily understood by individuals in informal sectors who may havelimited proficiency in official languages or foreign languages. This languagebarrier poses challenges in disseminating and implementing ethical principleseffectively. The informal sectors in Africa operate under unique economic dy-namics that may not align with formal economic structures. MI ethics guidelinescrafted for formal settings may not address the specific challenges, needs, andethical considerations prevalent in informal economies. Informal sectors oftenlack robust regulatory frameworks and oversight. The absence of clear regula-tions makes it difficult to enforce MI ethics guidelines, leaving informal workersand businesses vulnerable to ethical violations without proper accountabilitymechanisms. Awareness of MI ethics guidelines may be low in informal sectorsdue to limited access to information and educational resources. Individuals andbusinesses may not be aware of ethical considerations related to MI, hinderingthe adoption of ethical practices. Informal sectors face resource constraints,limiting their ability to invest in MI technologies that align with ethical guide-lines. Prioritizing ethical considerations may take a backseat when basic survivaland economic challenges are more pressing. Power imbalances within informalsectors, such as gender and socioeconomic disparities, may exacerbate ethicalchallenges. MI ethics guidelines may not sufficiently address these imbalances,leaving certain groups more vulnerable to negative impacts of MI technologies.Rapid changes in technology may outpace the adaptability of MI ethics guide-lines in informal sectors. Guidelines may struggle to keep pace with evolving MIapplications, leading to gaps in addressing emerging ethical issues. Historicaland societal factors may contribute to a trust deficit in certain informal sec-tors. MI ethics guidelines may face resistance or skepticism, requiring tailoredapproaches to build trust and ensure effective ethical practices.We look at a 10-row, 7-column table of ethics. We have selected only a fewtopics as rows as an illustration of context-sensibility of ethics and the topicscan be complemented. The selected rows are:• Community Engagement and Communal Values:• Cultural Sensitivity:• Environmental Stewardship:140• Social Justice and Equality:• Healthcare Equity:• Education Access:• Economic Inclusivity:• Technological Ethics:• Political Accountability:• Local, Inter-regional and International Collaboration.And the columns are:• Individual• Different Cultures• Institutions• Nation:• On and off-line platforms• Globally• Over Time.8.1 MI ethicsBelow we breakdown the technological ethics into 60 sub-items: 30 technical MIethics and 30 non-technical MI ethics in the African context. Each of the itemslayers by the 7-column multimodality.8.1.1 Technical MI ethics• Fair Algorithmic Design: Ensuring fairness in the technical design of al-gorithms to avoid biases and discriminatory outcomes.• Bias Detection and Mitigation: Developing techniques to identify andaddress biases within MI models.• Explainable MI: Creating MI systems that provide transparent explana-tions for their decision-making processes.• Privacy-Preserving Technologies: Implementing methods to protect indi-viduals’ privacy when handling sensitive data.• Robustness and Security: Designing MI systems resilient to adversarialattacks and ensuring cybersecurity measures.141• Accountability in Algorithmic Decision-Making: Establishing mechanismsto hold MI systems accountable for their decisions.• Bias-Free Data Collection: Ensuring the collection of diverse and repre-sentative datasets to reduce biases in MI models.• Ethical Data Governance: Implementing ethical frameworks for the col-lection, storage, and usage of data in MI applications.• Responsible MI Research: Conducting MI research with ethical consider-ations, transparency, and social impact in mind.• Model Fairness Evaluation: Developing metrics and standards to assessthe fairness of MI models.• Algorithmic Transparency: Making algorithms and decision processes un-derstandable to stakeholders.• Algorithmic Accountability: Assigning responsibility for the outcomes ofMI algorithms and decision-making.• Data Accuracy and Quality: Ensuring high-quality and accurate datainputs for MI models.• Human-Centric Design: Prioritizing the needs and well-being of humansin the design of MI systems.• Human-in-the-Loop Systems: Integrating human oversight and decision-making into MI processes.• Ethical Use of MI in Research: Ensuring that MI technologies are usedethically in academic and industrial research.• Continuous Monitoring and Auditing: Regularly assessing MI systems forethical compliance and performance.• Adversarial Robustness: Developing MI models that resist adversarial at-tempts to manipulate their behavior.• Fairness in Feature Engineering: Ensuring fairness in the selection anduse of features in MI models.• Secure Model Deployment: Implementing secure and ethical practiceswhen deploying MI models.• Algorithmic Impact Assessment: Evaluating the potential impact of algo-rithms on various stakeholders.• Interpretability of Predictions: Making MI predictions interpretable andunderstandable to end-users.142• Bias-Aware Machine Learning: Incorporating awareness of bias in machinelearning processes.• MI System Reliability: Ensuring the reliability and consistency of MIsystems in diverse scenarios.• Energy Efficiency: Developing MI models and algorithms that are envi-ronmentally sustainable.• Safety Mechanisms in Autonomous Systems: Implementing safety featuresin AI-powered autonomous systems.• Model Update Ethics: Addressing ethical considerations when updatingMI models over time.• Ethical Use of MI in Healthcare: Ensuring responsible deployment of MItechnologies in medical settings.• Cross-Cultural MI Understanding: Adapting MI systems to understandand respect diverse cultural contexts.• Secure Federated Learning: Ensuring the security and privacy of federatedlearning processes.8.1.2 Non-technical MI ethics• Informed Consent: Ensuring individuals are adequately informed and con-sent to the use of MI technologies.• Ethical Governance: Establishing ethical guidelines and policies for theresponsible development and deployment of MI.• Social Impact Assessment: Evaluating the broader societal impact of MItechnologies before deployment.• Inclusivity and Diversity: Promoting diversity in MI development teamsand considering diverse perspectives in MI applications.• Global Collaboration on MI Ethics: Fostering international cooperationto address global MI ethical challenges.• Human Rights Protection: Safeguarding human rights in the developmentand deployment of MI systems.• Ethical Education and Awareness: Promoting awareness and educationon MI ethics for stakeholders and the general public.• Cultural Sensitivity: Adapting MI technologies to respect and align withdiverse cultural norms and values.• Dual-Use Technology Awareness: Recognizing the potential dual-use na-ture of MI technologies and implementing safeguards.143• Public Participation: Involving the public in decision-making processesrelated to MI technologies to ensure inclusivity and representation.• Ethical Impact on Employment: Addressing the ethical implications ofMI on employment and workforce dynamics.• Responsible MI Procurement: Considering ethical factors when procuringMI technologies for public or private use.• Community Engagement: Engaging with local communities to understandand address their concerns about MI technologies.• MI and Democratic Values: Ensuring that MI technologies align withdemocratic principles and values.• User Empowerment: Empowering users to have control over their dataand MI interactions.• Digital Divide Reduction: Implementing strategies to bridge the digitaldivide and ensure equitable access to MI benefits.• Fair Distribution of MI Benefits: Ensuring that the benefits of MI tech-nologies are distributed equitably across society.• Long-Term Environmental Sustainability: Assessing and mitigating theenvironmental impact of MI technologies.• Ethical Marketing and Advertising: Ensuring ethical practices in the mar-keting and advertising of MI products and services.• Responsible MI Journalism: Ethically reporting on MI developments andensuring accuracy and transparency.• MI and Accessibility: Designing MI systems that are accessible to indi-viduals with diverse abilities.• MI and Cultural Heritage Preservation: Safeguarding cultural heritage inthe development and deployment of MI technologies.• Human-MI Collaboration Ethics: Establishing ethical guidelines for col-laborative interactions between humans and MI systems.• MI Impact on Social Equality: Considering and addressing the potentialimpact of MI on social equality.• Algorithmic Accountability in Government: Ensuring accountability forMI use in governmental decision-making.• Elderly and Vulnerable Population Protection: Safeguarding the rightsand well-being of elderly and vulnerable populations in MI applications.144• Ethical MI Policy Development: Creating policies that consider the ethicalimplications of MI technologies.• Cross-Industry Collaboration: Encouraging collaboration between differ-ent industries to address common MI ethical challenges.• MI and Child Protection: Implementing measures to protect children’sprivacy and well-being in MI applications.• MI for Social Good: Promoting the use of MI technologies for positivesocial impact and humanitarian causes.Tables 55, 56 and 57 display some examples of technical MI ethics. Tables58, 59 and 60 display some examples of non-technical MI ethics.Individuals Cultures Developers Institutions Nation Platforms Globally Over TimeFair Algorith-mic DesignCulturaldiversityinfluencesalgorithmfairnessDeveloper re-sponsibility inethical MIInstitutionalpolicies guideethical MINational regu-lations on MIethicsPlatform-specific ethi-cal guidelinesGlobal collab-orations forethical MIEthical con-siderationsevolve overtimeBias De-tection andMitigationIdentifyingand correctingalgorithmicbiasesDevelopertools for biasdetectionInstitutionalstrategies forbias mitiga-tionNational ini-tiatives toaddress biasPlatform-specific biashandlingmechanismsGlobal effortsto mitigate bi-ases in MIEthical normsadapt as bi-ases changeExplainableMITransparentalgorithms foruser under-standingDeveloperemphasis onexplainablemodelsInstitutionalpush fortransparentMI systemsNational de-mand forunderstand-able MIPlatform com-mitment toexplainabilityGlobal stan-dards for MItransparencyEthical im-portance ofexplainabilitygrowsPrivacy-PreservingTechnologiesIndividualprivacy prior-itized in MIsystemsDeveloper im-plementationof privacy-preservingtechniquesInstitutionalsafeguards foruser privacyNational lawsprotectingindividualprivacyPlatform-specific pri-vacy measuresGlobal con-sensus onMI privacystandardsEthical con-cerns over MIand privacypersistRobustnessand SecurityEnsuring MIsystems arerobust againstadversarialattacksDeveloper fo-cus on build-ing secure MImodelsInstitutionalmeasures forMI systemrobustnessNationalsecurity regu-lations for MIapplicationsPlatform-specific secu-rity protocolsGlobal collab-oration for MIsecurity stan-dardsEthical impli-cations of MIvulnerabilitiesAccountabilityin AlgorithmicDecision-MakingHolding in-dividualsaccountablefor MI deci-sionsDeveloperresponsibilityin algorithmicchoicesInstitutionalframeworksfor MI deci-sion account-abilityNationalguidelines foraccountableMI decisionsPlatform-specific ac-countabilitymechanismsGlobal ef-forts foraccountableMI practicesEthical con-siderationsof decisionresponsibilityBias-FreeData Collec-tionEnsuring un-biased datacollectionpracticesDevelopercommitmentto unbiaseddata gatheringInstitutionalguidelines forbias-free datacollectionNational regu-lations on fairdata collec-tionPlatform-specific dataethics policiesGlobal con-sensus onbias-free datapracticesEthical impor-tance of unbi-ased data col-lectionEthical DataGovernanceEthical man-agement anduse of dataDeveloperadherence toethical datapracticesInstitutionalpolicies forethical datagovernanceNationalframeworksfor responsibledata usePlatform-specific datagovernancestandardsGlobal col-laboration onethical datamanagementEvolving eth-ical norms indata gover-nanceResponsibleMI ResearchEthical con-siderationsin machinelearning re-searchDevelopercommitmentto responsibleMI researchInstitutionalsupportfor ethicalresearch prac-ticesNational fund-ing for respon-sible MI re-searchPlatform-specific re-search ethicsguidelinesGlobal ini-tiatives forresponsibleMI researchEthical chal-lenges inadvancing MIresearchModel Fair-ness Evalua-tionEvaluatingmodels forfairness andimpartialityDevelopertools formodel fairnessassessmentInstitutionalframeworksfor modelfairness evalu-ationNational stan-dards for fairMI modelsPlatform-specific fair-ness evalua-tion practicesGlobal bench-marks formodel fairnessEthicalscrutiny inassessingmodel fairnessTable 55: Some Examples of Technical MI Ethics 18.2 Rows: Selected ethics topicsCommunity Engagement and Communal Values• Individual: Active participation and collaboration within local commu-nities, emphasizing shared values, traditional practices, and community145AlgorithmicTransparencyOpenness inalgorithmicprocessesDevelopercommitmentto transparentalgorithmsInstitutionalpush for al-gorithmictransparencyNational reg-ulations onalgorithmictransparencyPlatform-specific trans-parencymeasuresGlobal stan-dards foralgorithmicopennessEthical impor-tance of trans-parency in al-gorithmsAlgorithmicAccountabil-ityHolding al-gorithmsaccountablefor their im-pactDeveloperresponsibilityin algorithmicoutcomesInstitutionalframeworksfor algorith-mic account-abilityNationalguidelines foraccountablealgorithmsPlatform-specific ac-countabilitymeasuresGlobal ef-forts foralgorithmicresponsibilityEthical con-siderationsin holdingalgorithmsaccountableData Ac-curacy andQualityEnsuring ac-curacy andquality indata used forMIDeveloper fo-cus on accu-rate and high-quality dataInstitutionalstandards fordata accuracyNational reg-ulations ondata qualityfor MIPlatform-specific dataquality mea-suresGlobal con-sensus on dataaccuracy inMIEthical impli-cations of in-accurate dataHuman-Centric De-signDesigning MIsystems with afocus on hu-man needsDeveloperemphasison human-centered MIdesignInstitutionalencourage-ment ofhuman-centricMINationalsupportfor human-focused MIdevelopmentPlatform com-mitment tohuman-centricdesignGlobal ini-tiatives forMI systemsaligned withhuman valuesEthical con-siderations inhuman-centricMIHuman-in-the-LoopSystemsIntegratinghuman in-put in MIdecision-makingDeveloper im-plementationof human-in-the-loopsystemsInstitutionalsupport forhuman in-volvement inMI processesNationalguidelines forhuman-in-the-loop MIPlatform-specific mech-anisms forhuman-in-the-loop MIGlobal stan-dards forhuman-in-the-loop MIsystemsEthical im-portance ofhuman in-volvement inMIEthical Useof MI in Re-searchEthical con-siderations inthe use of MIin researchDevelopercommitmentto ethicallyusing MI inresearchInstitutionalguidelines forthe ethicaluse of MI inresearchNationalframeworksfor responsibleMI researchusePlatform-specific ethicsin MI researchpracticesGlobal initia-tives for ethi-cal use of MIin researchEvolving eth-ical norms inMI researchuseContinuousMonitoringand AuditingRegular moni-toring and au-diting of MIsystemsDeveloperpractices forcontinuousMI systemevaluationInstitutionalframeworksfor MI systemmonitoringNational reg-ulations oncontinuousMI systemauditingPlatform-specific mon-itoring andauditing mea-suresGlobal stan-dards for MIsystem evalu-ationEthical impli-cations of con-tinuous moni-toring and au-ditingAdversarialRobustnessBuilding MIsystems re-silient toadversarialattacksDeveloperfocus onadversarialrobustness inMI modelsInstitutionalmeasures foradversarialresilience inMINationalguidelines foradversarial ro-bustness in MIapplicationsPlatform-specific de-fenses againstadversarialattacksGlobal col-laboration foradversarialrobustness inMIEthical con-siderationsin addressingadversarialvulnerabilitiesFairness inFeature Engi-neeringEnsuring fair-ness in the se-lection of fea-tures for MImodelsDeveloper at-tention to fair-ness in featureengineeringInstitutionalguidelines forfair featureselectionNational regu-lations on fair-ness in featureengineeringPlatform-specific fair-ness consid-erations infeature engi-neeringGlobal con-sensus onfairness inMI featureselectionEthical im-plications ofbiased featureengineeringSecure ModelDeploymentEnsuringsecure deploy-ment of MImodelsDeveloperpractices forsecure modeldeploymentInstitutionalmeasures forsecure MImodel deploy-mentNationalguidelines forsecure MImodel deploy-mentPlatform-specific secu-rity protocolsfor modeldeploymentGlobal stan-dards forsecure MImodel deploy-mentEthical con-siderations insecure modeldeploymentTable 56: Some Examples of Technical MI Ethics 2development.• Different Cultures: Respect for diverse ethnic, linguistic, and culturalidentities, fostering unity and cooperation while preserving and celebratingunique cultural expressions.• Institutions: Support for community-based institutions that uphold com-munal values, ensuring their integration into broader societal frameworksand decision-making processes.• Nation: National policies promoting community engagement, preservingcultural heritage, and recognizing the importance of communal values inshaping the nation’s identity.• Globally: Collaboration with global entities while maintaining a strongconnection to African cultural values, contributing to a more inclusiveand culturally sensitive global community.146AlgorithmicImpact As-sessmentAssessing so-cietal impactof algorithmsDevelopertools forimpact assess-mentInstitutionalframeworksfor algorith-mic impactevaluationNationalguidelines foralgorithmicimpact assess-mentPlatform-specificimpact as-sessmentpracticesGlobal stan-dards foralgorith-mic societalimpact assess-mentEthical im-plicationsin assessingalgorithmicimpactInterpretabilityof PredictionsMaking MIpredictionsinterpretableDeveloperfocus on in-terpretablepredictionmodelsInstitutionalencourage-ment ofinterpretableMI predictionsNationalsupport forinterpretableMI predictionsPlatform-specific in-terpretablepredictionmechanismsGlobal stan-dards forinterpretableMI predictionsEthical con-siderations inprediction in-terpretabilityBias-AwareMachineLearningDeveloping MImodels withawareness ofbiasesDeveloperemphasis onbias-aware MLmodelsInstitutionalguidelines forbias-awaremachinelearningNational reg-ulations onbias-awaremachinelearningPlatform-specific mech-anisms foraddressingbiases in MLGlobal col-laboration forbias-aware MLpracticesEthical con-siderations indeveloping MImodels withbias awarenessMI System Re-liabilityEnsuring reli-ability in MIsystemsDeveloperpractices forreliable MImodelsInstitutionalmeasures forMI systemreliabilityNational stan-dards for reli-able MI appli-cationsPlatform-specificreliabilityassuranceGlobal stan-dards for MIsystem relia-bilityEthical im-plications ofunreliable MIsystemsEnergy Effi-ciencyDevelopingenergy-efficientMI solutionsDeveloperfocus on MImodels withlow energyconsumptionInstitutionalencour-agementof energy-efficient MINationalsupport foreco-friendlyMI technolo-giesPlatform-specific en-ergy efficiencymeasuresGlobal ini-tiativesfor energy-efficient MIEthical con-siderations inMI’s environ-mental impactSafety Mecha-nisms in Au-tonomous Sys-temsImplementingsafety mea-sures inautonomousMI systemsDeveloper in-corporation ofsafety mecha-nisms in au-tonomyInstitutionalframeworksfor safety inautonomoussystemsNational reg-ulations onsafety in au-tonomous MIPlatform-specific safetyprotocols forautonomyGlobal col-laborationon safetystandards inautonomousMIEthical im-plicationsof safety inautonomoussystemsModel UpdateEthicsEthical con-siderations inupdating MImodelsDevelopercommitmentto ethicalmodel updatesInstitutionalguidelines forethical modelupdatesNationalframeworksfor responsiblemodel updatesPlatform-specific ethi-cal considera-tions in modelupdatesGlobal initia-tives for ethi-cal model up-datesEvolving eth-ical norms inupdating MImodelsEthical Use ofMI in Health-careEthical con-siderations indeploying MIin healthcareDevelopercommitmentto ethicaluse of MI inhealth appli-cationsInstitutionalguidelinesfor ethicaluse of MI inhealthcareNational reg-ulations onethical MI inhealthcarePlatform-specific ethicsin MI forhealthcareGlobal initia-tives for ethi-cal use of MIin healthcareEvolving eth-ical norms inMI’s role inhealthcareCross-CulturalMI Under-standingEnsuringcultural un-derstanding inMI modelsDeveloperfocus oncross-culturalMI modelsInstitutionalsupport forcross-culturalMI under-standingNational en-couragementof culturally-aware MIPlatform-specific mea-sures forcross-culturalMIGlobal col-laboration forculturally-sensitive MIEthical con-siderations incross-culturalMI under-standingSecure Feder-ated LearningEnsuring secu-rity in feder-ated learningsetupsDeveloperpractices forsecure feder-ated learningInstitutionalmeasures forsecure feder-ated MINationalguidelines forsecure feder-ated learningPlatform-specific secu-rity protocolsfor federatedMIGlobal stan-dards forsecure feder-ated learningEthical impli-cations of se-curity in fed-erated MITable 57: Some examples of Technical MI Ethics 3• Over Time: Preservation and adaptation of communal values over genera-tions, acknowledging their historical significance and evolving in harmonywith changing societal needs.Cultural Sensitivity• Individual: Personal awareness and respect for the rich tapestry of Africancultures, promoting inclusive attitudes and fostering cross-cultural under-standing in daily interactions.• Different Cultures: Encouragement of cultural exchange, dialogue, andmutual appreciation among diverse African cultures, emphasizing the im-portance of preserving cultural heritage.• Institutions: Integration of cultural sensitivity into institutional policies,ensuring fair representation and equitable treatment of individuals fromvarious cultural backgrounds.147Individuals Cultures Developers Institutions Nation Platforms Globally Over TimeInformed Con-sentIndividualsmaking in-formed deci-sions in MIuseDevelopercommunica-tion of MIimplicationsInstitutionalpractices forinformed con-sentNational lawson informedconsent for MIPlatform-specific con-sent mecha-nismsGlobal stan-dards for MIinformed con-sentChangingnorms inMI informeddecision-makingEthical Gover-nanceEthicaldecision-making atgovernancelevelsDeveloperadherenceto ethicalgovernanceprinciplesInstitutionalframeworksfor ethicalgovernanceNational poli-cies on MIethical gover-nancePlatform-specific ethi-cal governancestructuresGlobal collab-oration on MIethical gover-nanceEvolving ethi-cal governancenormsSocial ImpactAssessmentAssessing so-cietal impactof MI imple-mentationsDevelopertools for so-cial impactassessmentInstitutionalframeworksfor MI socialimpact assess-mentNationalguidelinesfor assessingMI’s societalimpactPlatform-specificimpact as-sessmentpracticesGlobal stan-dards for MIsocial impactassessmentEthical con-siderationsin assessingMI’s societalimpactInclusivityand DiversityEnsuring in-clusivity anddiversity inMI useDeveloper fo-cus on inclu-sive and di-verse MI mod-elsInstitutionalsupport forinclusive MIpracticesNational en-couragementof diverse MIapplicationsPlatform-specific mea-sures forinclusivityand diversityGlobal collab-oration for di-verse MIEthical con-siderationsin promotinginclusivityand diversityGlobal Collab-oration on MIEthicsCollaboratingglobally onethical MIstandardsDeveloperengagementin global MIethics discus-sionsInstitutionalsupport forinternationalMI ethicscollaborationsNational in-volvementin global MIethics initia-tivesPlatform-specific globalcollaborationmechanismsWorldwidecooperationon MI ethicsEvolvingglobal ethicalnorms in MIHuman RightsProtectionProtecting hu-man rights inMI useDevelopercommitmentto humanrights in MIapplicationsInstitutionalsafeguards forMI and humanrightsNational lawson MI andhuman rightsprotectionPlatform-specific hu-man rightsconsiderationsGlobal stan-dards for MIand humanrightsEthical con-siderationsin protectinghuman rightsin MIEthical Ed-ucation andAwarenessPromotingethical MIeducationDeveloper em-phasis on eth-ical MI educa-tionInstitutionalsupport forMI ethicseducationNational ini-tiatives forethical MIeducationPlatform-specific ed-ucationalprograms onMI ethicsGlobal effortsfor MI ethicseducationEvolving eth-ical educationnorms in MICultural Sen-sitivityEnsuring cul-tural sensitiv-ity in MI mod-elsDeveloperfocus onculturally-sensitive MImodelsInstitutionalencourage-ment ofculturally-aware MINationalsupport forculturally-sensitive MIapplicationsPlatform-specificmeasuresfor culturalsensitivity inMIGlobal col-laboration forculturally-sensitive MIEthical con-siderationsin culturalsensitivity inMIDual-UseTechnologyAwarenessRaising aware-ness of dual-use MI tech-nologiesDeveloperconsiderationof dual-useimplicationsInstitutionalframeworksfor dual-useMI awarenessNational reg-ulations ondual-use MIapplicationsPlatform-specific dual-use technologyguidelinesGlobal col-laboration ondual-use MIawarenessEthical im-plications ofdual-use MItechnologiesPublic Partici-pationInvolvingthe public inMI decision-makingDeveloperpractices forpublic in-volvement inMIInstitutionalframeworksfor publicparticipationin MINationalguidelines forpublic engage-ment in MIPlatform-specific mech-anisms forpublic in-volvementGlobal stan-dards forpublic partici-pation in MIEthical con-siderations inpublic engage-ment in MITable 58: Some Examples of Non-Technical MI Ethics 1• Nation: National efforts to protect and promote cultural diversity, recog-nizing the intrinsic value of each culture in contributing to the nation’sidentity.• Globally: Advocacy for global recognition and respect of African cultures,challenging stereotypes and fostering a positive image of Africa on theglobal stage.• Over Time: Preservation of cultural heritage and adaptation to contem-porary contexts, allowing cultural traditions to evolve while maintainingtheir authenticity.Environmental Stewardship• Individual: Sustainable practices and environmental awareness at the in-dividual level, emphasizing the importance of preserving Africa’s diverseecosystems and natural resources.148Ethical Im-pact onEmploymentConsideringethical impacton employ-ment in MIDeveloperawareness ofemploymentimplicationsInstitutionalframeworksfor ethicalemploymentimpactNational poli-cies on MI’sethical impacton employ-mentPlatform-specific con-siderationsfor ethicalemploymentimpactGlobal col-laborationon ethicalemploymentimpactEvolving eth-ical norms inMI’s impacton employ-mentResponsibleMI Procure-mentEthical con-siderationsin MI systemprocurementDevelopercommitmentto responsibleMI procure-mentInstitutionalguidelines forethical MIprocurementNational stan-dards forresponsibleMI systemacquisitionPlatform-specific con-siderations inMI procure-mentGlobal ini-tiatives forresponsibleMI procure-mentEvolvingnorms inethical MIprocurementCommunityEngagementEngaging withcommunitiesin MI develop-mentDeveloperpractices forcommunityengagement inMIInstitutionalframeworksfor communityinvolvementin MINationalpolicies oncommunityengagement inMIPlatform-specificcommunityengagementmechanismsGlobal stan-dards for MIcommunityinvolvementEthical con-siderations incommunityengagementMI and Demo-cratic ValuesUpholdingdemocraticvalues in MIapplicationsDeveloperadherence todemocraticprinciples inMIInstitutionalsupport fordemocratic MIpracticesNational lawson MI anddemocraticvaluesPlatform-specific demo-cratic MIconsiderationsGlobal col-laboration ondemocratic MIstandardsEvolvingnorms in MI’salignmentwith demo-cratic valuesUser Empow-ermentEmpoweringusers in MIinteractionsDeveloper fo-cus on empow-ering users inMI systemsInstitutionalsupport foruser empower-ment in MINationalguidelines foruser empower-ment in MIPlatform-specific mech-anisms foruser empower-mentGlobal stan-dards for MIuser empower-mentEthical con-siderations inuser empower-ment in MIDigital DivideReductionReducing dig-ital divideswith ethicalMIDeveloperefforts to ad-dress digitalinequalitiesInstitutionalstrategies forreducing thedigital divideNational poli-cies on digitaldivide reduc-tion throughMIPlatform-specific mea-sures fordigital dividereductionGlobal collab-oration on MIfor digital di-vide reductionEthical impli-cations of MIin addressingdigital dispar-itiesFair Distri-bution of MIBenefitsEnsuring fairdistribution ofMI benefitsDeveloperconsiderationof benefitdistribution inMIInstitutionalframeworksfor fair benefitdistributionNational poli-cies on equi-table MI bene-fitsPlatform-specific mech-anisms for fairMI benefitdistributionGlobal col-laboration forequitable MIbenefitsEthical con-siderationsin MI benefitdistributionLong-TermEnvironmen-tal Sustain-abilityConsideringlong-termenvironmentalimpact in MIDeveloperawarenessof MI’s en-vironmentalfootprintInstitutionalframeworksfor sustainableMI practicesNational reg-ulations onMI and en-vironmentalsustainabilityPlatform-specific mea-sures for en-vironmentallysustainableMIGlobal col-laboration onMI’s environ-mental impactEthical im-plications ofMI’s envi-ronmentalfootprintEthical Mar-keting andAdvertisingEthical con-siderations inMI marketingDeveloperpractices forethical MIadvertisingInstitutionalguidelines forethical MImarketingNational stan-dards forresponsibleMI promotionPlatform-specific ethi-cal marketingin MIGlobal initia-tives for ethi-cal MI adver-tisingEvolvingnorms in MImarketingethicsResponsibleMI JournalismEthical con-siderations inreporting onMIDevelopercommitmentto responsibleMI journalismInstitutionalguidelines forethical report-ing on MINational stan-dards for jour-nalism on MIPlatform-specific re-sponsible MIreportingpracticesGlobal col-laboration onethical MIjournalismEvolvingnorms in MIjournalismethicsTable 59: Some examples of Non-Technical MI Ethics 2• Different Cultures: Incorporation of traditional ecological knowledge intoenvironmental conservation efforts, recognizing the symbiotic relationshipbetween African cultures and nature.• Institutions: Implementation of policies that prioritize environmental sus-tainability, considering the impact on local ecosystems and ensuring re-sponsible resource management.• Nation: National strategies for environmental protection, addressing chal-lenges like deforestation, wildlife conservation, and climate change adap-tation in the African context.• Globally: Collaboration on global environmental initiatives, advocatingfor equitable representation and acknowledging Africa’s role in the globalecological balance.• Over Time: Long-term environmental planning, balancing developmentwith conservation to secure a sustainable and thriving environment for149MI and Acces-sibilityEnsuring ac-cessibility inMI applica-tionsDeveloper fo-cus on accessi-ble MI modelsInstitutionalsupport foraccessible MIpracticesNational regu-lations on MIaccessibilityPlatform-specific mea-sures for MIaccessibilityGlobal collab-oration for ac-cessible MIEthical con-siderations inMI accessibil-ityMI and Cul-tural HeritagePreservationEthical con-siderationsin preservingcultural her-itage with MIDeveloperpractices forculturally-sensitive MIin heritagepreservationInstitutionalframeworksfor ethicalcultural her-itage MINationalguidelines forpreservingcultural her-itage with MIPlatform-specific cul-tural heritageMI considera-tionsGlobal col-laboration onMI for cul-tural heritagepreservationEvolvingnorms in MI’srole in cul-tural heritageHuman-MICollaborationEthicsEthical con-siderationsin human-MIcollaborationsDevelopercommitmentto ethicalhuman-MIcollaborationInstitutionalguidelinesfor ethicalhuman-MIpartnershipsNational poli-cies on ethicalMI collabora-tion with hu-mansPlatform-specific ethicsin human-MIcooperationGlobal stan-dards for ethi-cal human-MIcollaborationEvolvingnorms in MI’sinteractionwith humansMI Impact onSocial Equal-ityConsideringsocial equalityin MI impactDeveloperawareness ofMI’s impacton socialequalityInstitutionalframeworksfor assessingMI’s impacton socialequalityNational poli-cies on MI andsocial equalityPlatform-specific mea-sures forpromotingsocial equalityin MIGlobal col-laboration onMI’s impacton socialequalityEthical im-plications ofMI’s influ-ence on socialequalityAlgorithmicAccount-ability inGovernmentHolding gov-ernmentalgorithmsaccountableDeveloperpractices foraccountablegovernmentalgorithmsInstitutionalframeworksfor algorith-mic account-ability ingovernmentNational regu-lations on MIaccountabilityin governmentPlatform-specific mech-anisms foraccountablegovernmentalgorithmsGlobal collab-oration on ac-countable gov-ernment MIEthical con-siderations ingovernmentalgorithmaccountabilityElderly andVulnerablePopulationProtectionProtecting el-derly and vul-nerable popu-lations in MIDeveloperconsiderationof MI’s impacton vulnerablepopulationsInstitutionalsafeguards forelderly andvulnerableindividuals inMI useNationalguidelines forprotectingvulnerablepopulationswith MIPlatform-specific mea-sures forsafeguardingvulnerablepopulationsGlobal collab-oration on MIprotection forvulnerable in-dividualsEthical im-plications ofMI’s influenceon elderly andvulnerablepopulationsEthical MIPolicy Devel-opmentEthical con-siderationsin MI policycreationDevelopercommitmentto ethical MIpolicy devel-opmentInstitutionalguidelines forethical MIpolicy formu-lationNational poli-cies on creat-ing ethical MIframeworksPlatform-specific ethi-cal MI policydevelopmentGlobal col-laborationon ethicalMI policycreationEvolvingnorms in MIpolicy devel-opment ethicsCross-IndustryCollaborationand Coopeti-tionCollaboratingacross in-dustries forethical MIDeveloperengagement incross-industryMI collabora-tionsInstitutionalsupport forethical collab-oration andcoopetition inMINational en-couragementof cross-industry MIpartnershipsPlatform-specific mech-anisms forcross-industryMI collabora-tionGlobal stan-dards forcross-industryMI coopeti-tionEthical impli-cations of MIcollaborationacross indus-triesMI and ChildProtectionEnsuring childprotection inMI applica-tionsDeveloper fo-cus on child-safe MI mod-elsInstitutionalsupport forMI applica-tions ensuringchild safetyNational regu-lations on MIand child pro-tectionPlatform-specific mea-sures forchild-safe MIGlobal collab-oration on MIfor child pro-tectionEthical con-siderations inMI’s impacton child safetyMI for SocialGoodPromoting MIfor positivesocial impactDevelopercommitmentto using MIfor social goodInstitutionalframeworksfor MI ap-plicationsbenefitingsocietyNational poli-cies on MI forsocial goodPlatform-specific initia-tives for MI’spositive socialimpactGlobal collab-oration on MIfor social goodEvolvingnorms in MI’srole for soci-etal benefitTable 60: Some examples of Non-Technical MI Ethics 3future generations.Social Justice and Equality• Individual: Advocacy for social justice, equality, and human rights at theindividual level, addressing issues like gender inequality, discrimination,and social disparities.• Different Cultures: Recognition and celebration of diversity, promotinginclusive social structures that respect and uphold the rights of individualsfrom various cultural backgrounds.• Institutions: Development and implementation of policies that combat so-cial injustices, ensuring equal opportunities and fair representation withinsocietal institutions.150• Nation: National commitment to social justice, actively addressing his-torical inequalities and working towards a more inclusive and equitablesociety.• Globally: Participation in global movements for social justice, challengingsystemic inequalities on the international stage and advocating for Africa’sinterests.• Over Time: Ongoing efforts to eradicate social injustices, acknowledginghistorical challenges and working towards a more just and equal future.Healthcare Equity• Individual: Access to healthcare education and proactive engagement inpersonal health practices, promoting preventive measures and awarenesswithin local communities.• Different Cultures: Integration of culturally sensitive healthcare practices,recognizing traditional healing methods and addressing healthcare dispar-ities in diverse cultural contexts.• Institutions: Development of healthcare policies that prioritize equity, en-suring equal access to quality healthcare services for all citizens.• Nation: National healthcare strategies that address public health chal-lenges, emphasizing universal access to healthcare and reducing healthinequalities.• Globally: Collaboration on global health initiatives, advocating for fairdistribution of resources and acknowledging Africa’s contributions to globalhealth.• Over Time: Progress in healthcare infrastructure and services, adapting toevolving health needs and ensuring continuous improvement in healthcareequity.Education Access• Individual: Pursuit of education and lifelong learning, emphasizing thevalue of knowledge and skills to empower individuals and contribute tocommunity development.• Different Cultures: Recognition of diverse learning styles and educationalapproaches, fostering an inclusive education system that respects variouscultural perspectives.• Institutions: Implementation of inclusive educational policies, ensuringequal access to quality education for individuals from different culturalbackgrounds.151• Nation: National commitment to education as a key driver of develop-ment, investing in educational infrastructure and promoting access for allcitizens.• Globally: Engagement in international collaborations on educational ini-tiatives, contributing African perspectives to global education discussions.• Over Time: Evolution of education systems to meet changing needs,adapting to technological advancements and ensuring education remainsa cornerstone of societal progress.Economic Inclusivity• Individual: Entrepreneurship, financial literacy, and economic empower-ment at the individual level, fostering economic independence and com-munity development.• Different Cultures: Recognition and integration of diverse economic mod-els, acknowledging the richness of various African economic practices andpromoting inclusive economic policies.• Institutions: Implementation of policies that promote economic inclusiv-ity, ensuring fair distribution of resources and opportunities within thenational economic framework.• Nation: National strategies for inclusive economic development, address-ing economic disparities and fostering economic resilience.• Globally: Engagement in global economic collaborations, advocating forfair representation and acknowledging Africa’s potential in shaping globaleconomic landscapes.• Over Time: Adaptation of economic strategies to evolving global markets,balancing traditional economic practices with innovative approaches forsustainable economic inclusivity.Technological Ethics• Individual: Ethical use of technology, digital literacy, and participation intechnological advancements, recognizing the role of technology in shapingthe African narrative.• Different Cultures: Integration of diverse perspectives into technologicalinnovation, ensuring that technological solutions consider the cultural con-text and benefit diverse communities.• Institutions: Development of ethical guidelines for technology use, foster-ing responsible innovation and addressing potential ethical challenges intechnological advancements.152• Nation: National policies that promote responsible and inclusive technol-ogy development, bridging the digital divide and ensuring equitable accessto technological benefits.• Globally: Participation in global discussions on technological ethics, ad-vocating for fair representation and ethical considerations that addressAfrica’s unique challenges and opportunities.• Over Time: Ethical adaptation to technological advancements, recogniz-ing the evolving role of technology in African societies and ensuring itsalignment with ethical principles over time.Political Accountability• Individual: Civic engagement, political awareness, and active participa-tion at the individual level, fostering a sense of responsibility and account-ability in governance.• Different Cultures: Recognition of diverse political perspectives, promot-ing inclusive political discourse that considers cultural nuances and diversevoices.• Institutions: Development of transparent and accountable political sys-tems, ensuring representation and responsiveness to the needs of diversecultural and social groups.• Nation: National commitment to political accountability, addressing cor-ruption, promoting transparency, and ensuring fair and inclusive politicalprocesses that respect the diverse cultural and social fabric of the nation.• Globally: Active engagement in international political collaborations, con-tributing African perspectives to global governance discussions, and ad-vocating for fair representation on the global stage.• Over Time: Evolution of political systems that adapt to changing soci-etal needs, acknowledging historical challenges, and continuously workingtowards accountable and inclusive governance over time.Local, Inter-regional, and International Collaboration• Individual: Promotion of collaboration at the local level, emphasizingcommunity engagement and fostering partnerships for community devel-opment.• Different Cultures: Integration of diverse cultural perspectives in collabo-rative initiatives, recognizing the richness of cultural diversity in shapingcollaborative efforts.153• Institutions: Development of policies that facilitate inter-regional andinternational collaboration, ensuring equitable participation and mutualbenefit.• Nation: National strategies for diplomatic collaboration, promoting Africa’sinterests on the international stage while respecting the diversity of Africannations.• Globally: Active engagement in global collaborations, contributing Africanperspectives to international discussions and fostering partnerships thataddress global challenges.• Over Time: Adaptation of collaboration strategies to changing global dy-namics, recognizing historical collaborations and continuously working to-wards more effective and inclusive collaborative efforts over time.In the African context, each of these rows emphasizes the unique challenges,opportunities, and values that contribute to the multifaceted MI ethical land-scape. The interconnectedness of these aspects reflects the dynamic and evolvingnature of MI ethics in Africa, where the preservation of cultural identity, envi-ronmental sustainability, social justice, and inclusive economic development areintegral components of MI ethical considerations across various scales and overtime.8.3 Columns: MI Ethics from individuals to the globelevelIndividual• Community Engagement and Communal Values: Personal commitment toactively participate in and contribute to local communities, emphasizingshared values and collaborative initiatives for community development.• Cultural Sensitivity: Personal awareness and respect for the diverse cul-tures within Africa, fostering an inclusive attitude that appreciates andcelebrates the richness of cultural identities.• Environmental Stewardship: Individual responsibility for sustainable prac-tices, including ecological awareness and a commitment to preservingAfrica’s diverse ecosystems and natural resources.• Social Justice and Equality: Personal advocacy for social justice, humanrights, and equality, addressing issues like gender inequality, discrimina-tion, and socioeconomic disparities.• Healthcare Equity: Commitment to personal health practices and proac-tive engagement in healthcare education, contributing to equitable accessto healthcare services.154• Education Access: Pursuit of education and lifelong learning, recognizingthe transformative power of education in personal and community devel-opment.• Economic Inclusivity: Individual initiatives in entrepreneurship, financialliteracy, and economic empowerment, fostering economic independenceand community resilience.• Technological Ethics: Ethical use of technology, digital literacy, and re-sponsible participation in technological advancements, considering the im-pact of technology on African societies.• Political Accountability: Civic engagement, political awareness, and ac-tive participation at the individual level, contributing to transparent andaccountable political processes.• Local, Inter-regional, and International Collaboration: Promotion of col-laboration at the local level, emphasizing community engagement and fos-tering partnerships for community development.Different Cultures• Community Engagement and Communal Values: Embracing and integrat-ing diverse cultural practices within local communities, promoting unitywhile respecting unique cultural expressions.• Cultural Sensitivity: Encouraging cultural exchange, dialogue, and mu-tual appreciation among diverse African cultures, fostering cross-culturalunderstanding and harmony.• Environmental Stewardship: Incorporating traditional ecological knowl-edge into environmental conservation efforts, recognizing the symbioticrelationship between African cultures and nature.• Social Justice and Equality: Recognition and celebration of diversity, pro-moting inclusive social structures that respect and uphold the rights ofindividuals from various cultural backgrounds.• Healthcare Equity: Integration of culturally sensitive healthcare practices,acknowledging traditional healing methods and addressing healthcare dis-parities in diverse cultural contexts.• Education Access: Recognition of diverse learning styles and educationalapproaches, fostering an inclusive education system that respects variouscultural perspectives.• Economic Inclusivity: Acknowledgment and integration of diverse eco-nomic models, ensuring that economic policies consider the richness ofvarious African economic practices.155• Technological Ethics: Integration of diverse perspectives into technologi-cal innovation, ensuring that technological solutions consider the culturalcontext and benefit diverse communities.• Political Accountability: Recognition of diverse political perspectives, pro-moting inclusive political discourse that considers cultural nuances anddiverse voices.• Local, Inter-regional, and International Collaboration: Integration of di-verse cultural perspectives in collaborative initiatives, recognizing the rich-ness of cultural diversity in shaping collaborative efforts.Institutions• Community Engagement and Communal Values: Support for community-based institutions that uphold communal values, ensuring their integrationinto broader societal frameworks and decision-making processes.• Cultural Sensitivity: Integration of cultural sensitivity into institutionalpolicies, ensuring fair representation and equitable treatment of individu-als from various cultural backgrounds.• Environmental Stewardship: Implementation of policies that prioritize en-vironmental sustainability, considering the impact on local ecosystems andensuring responsible resource management.• Social Justice and Equality: Development and implementation of poli-cies that combat social injustices, ensuring equal opportunities and fairrepresentation within societal institutions.• Healthcare Equity: Development of healthcare policies that prioritize eq-uity, ensuring equal access to quality healthcare services for all citizens.• Education Access: Implementation of inclusive educational policies, ensur-ing equal access to quality education for individuals from different culturalbackgrounds.• Economic Inclusivity: Implementation of policies that promote economicinclusivity, ensuring fair distribution of resources and opportunities withinthe national economic framework.• Technological Ethics: Development of ethical guidelines for technologyuse, fostering responsible innovation and addressing potential ethical chal-lenges in technological advancements.• Political Accountability: Development of transparent and accountable po-litical systems, ensuring representation and responsiveness to the needs ofdiverse cultural and social groups.156• Local, Inter-regional, and International Collaboration: Development ofpolicies that facilitate inter-regional and international collaboration, en-suring equitable participation and mutual benefit.Nation• Community Engagement and Communal Values: National policies pro-moting community engagement, preserving cultural heritage, and recog-nizing the importance of communal values in shaping the nation’s identity.• Cultural Sensitivity: National efforts to protect and promote cultural di-versity, recognizing the intrinsic value of each culture in contributing tothe nation’s identity.• Environmental Stewardship: National strategies for environmental pro-tection, addressing challenges like deforestation, wildlife conservation, andclimate change adaptation in the African context.• Social Justice and Equality: Policies should aim to address social injus-tices and promote equality, ensuring that MI benefits are accessible to allsegments of society.• Healthcare Equity: National MI in healthcare should focus on equity,providing accessible and affordable healthcare solutions for all citizens.• Education Access: Policies should promote equal access to educationthrough MI, bridging educational gaps and ensuring inclusivity.• Economic Inclusivity: National MI strategies should foster economic in-clusivity, preventing disparities and ensuring that the economic benefitsof MI are widespread.• Technological Ethics: National policies should address ethical considera-tions related to technology, promoting responsible and ethical MI prac-tices.• Political Accountability: Policies should incorporate mechanisms for po-litical accountability, ensuring that MI decisions align with democraticvalues and are subject to scrutiny.• Local, Inter-regional and International Collaboration: Collaboration atdifferent levels is essential, involving local communities, inter-regionalpartnerships, and international collaborations to address global challengesand share best practices.On and Off-Line Platforms• Community Engagement and Communal Values: Online platforms shouldengage communities, respecting communal values, and ensuring that MI-interactions align with local norms.157• Cultural Sensitivity: Platforms should be culturally sensitive, catering todiverse cultural backgrounds and ensuring that online MI experiences arerespectful.• Environmental Stewardship: Platforms should consider environmental im-pacts, adopting eco-friendly practices in data centers and technology in-frastructure.• Social Justice and Equality: Online platforms should promote social jus-tice and equality, preventing discrimination and ensuring fair representa-tion.• Healthcare Equity: Online healthcare platforms should prioritize equity,providing accessible healthcare information and services to all users.• Education Access: Online education platforms should ensure equal accessto educational resources, addressing digital divides and promoting inclu-sivity.• Economic Inclusivity: Platforms should promote economic inclusivity, pre-venting digital exclusion and ensuring equal opportunities for online par-ticipation.• Technological Ethics: Online platforms should adhere to ethical standards,ensuring responsible data handling, privacy protection, and transparentalgorithms.• Political Accountability: Platforms should establish mechanisms for po-litical accountability, ensuring transparent governance and accountabilityin MI-driven decisions.• Local, Inter-regional and International Collaboration: Collaboration onplatforms should extend beyond borders, involving local, inter-regional,and international partnerships to address global challenges and ensurediverse perspectives.Globally• Community Engagement and Communal Values: Global MI initiativesshould engage communities worldwide, respecting diverse communal val-ues and incorporating them into decision-making processes.• Cultural Sensitivity: Global collaborations should be culturally sensitive,considering and respecting diverse cultural values across different regions.• Environmental Stewardship: Global MI efforts should prioritize environ-mental stewardship, promoting sustainable practices and minimizing thecarbon footprint of MI technologies.158• Social Justice and Equality: Global MI collaborations should addressglobal social injustices and promote equality on a worldwide scale.• Healthcare Equity: Global efforts in MI healthcare should focus on eq-uity, ensuring that healthcare solutions are accessible and beneficial topopulations globally.• Education Access: Global initiatives should promote equal access to edu-cation through MI, bridging educational gaps and ensuring global educa-tional inclusivity.• Economic Inclusivity: Global MI strategies should foster economic inclu-sivity, preventing global disparities and ensuring that economic benefitsare distributed equitably.• Technological Ethics: Global collaborations should uphold technologicalethics, establishing common ethical standards and principles for MI de-velopment and deployment globally.• Political Accountability: Global initiatives should incorporate mechanismsfor political accountability, ensuring that MI decisions align with demo-cratic values and are subject to international scrutiny.• Local, Inter-regional and International Collaboration: Collaboration at aglobal level is essential, involving local, inter-regional, and internationalpartnerships to address global challenges and share best practices on abroader scale.Over Time• Community Engagement and Communal Values: Changes over time shouldinvolve ongoing community engagement, ensuring that communal valuesare considered and respected in evolving MI strategies.• Cultural Sensitivity: Evolving MI strategies should remain culturally sen-sitive, adapting to changing cultural contexts over time.• Environmental Stewardship: MI developments should adapt over time toalign with evolving environmental standards and best practices.• Social Justice and Equality: Over time, MI policies should adapt to ad-dress emerging social justice issues and promote ongoing equality.• Healthcare Equity: MI in healthcare should adapt over time to ensureongoing equity, addressing evolving healthcare challenges and disparities.• Education Access: MI in education should evolve over time to meet chang-ing educational needs and ensure continuous access to educational re-sources.159• Economic Inclusivity: Over time, MI strategies should adapt to fosterongoing economic inclusivity, preventing the emergence of new disparities.• Technological Ethics: Over time, technological ethics should evolve toaddress emerging ethical challenges associated with advancing MI tech-nologies.• Political Accountability: Over time, MI policies should adapt to maintainpolitical accountability, addressing new challenges and ensuring ongoingtransparency in political decisions influenced by MI.• Local, Inter-regional and International Collaboration: Collaboration shouldcontinue over time, adapting to changing global dynamics and ensuringthat MI efforts remain collaborative, inclusive, and responsive to evolvingchallenges.8.4 Multi-scale ethics interactionsThe strong interactions between the rows and columns of ethics in the multiscaleframework create a comprehensive and interconnected ethical landscape. Let’sexplore some of their relationships:Community Engagement and Communal Values• Individual: Transparency and fairness at the individual level foster a senseof responsibility and equity in engaging with communal values.• Different Cultures: Cultural sensitivity ensures that diverse cultural per-spectives are respected in communal discussions.• Institutions: Institutional responsibility shapes the transparency and fair-ness of communal engagements.• Nation: Political accountability contributes to the responsible handling ofcommunal values on a national scale.• Globally: International collaboration promotes a shared understanding ofcommunal values globally.• Over Time: Historical perspectives guide the context-awareness and equi-table evolution of communal values.Cultural Sensitivity• Individual: Personal awareness and responsibility enhance cultural sensi-tivity in individual interactions.• Different Cultures: Collaborative efforts and fairness promote culturalsensitivity among diverse cultural groups.160• Institutions: Institutional responsibility ensures cultural sensitivity in poli-cies and practices.• Nation: Political accountability addresses cultural biases, fostering cul-tural sensitivity on a national level.• Globally: Global collaboration shapes cultural sensitivity on a worldwidescale.• Over Time: Historical perspectives contribute to the context-awarenessand equitable evolution of cultural sensitivity.Environmental Stewardship• Individual: Responsible and transparent individual actions contribute toenvironmental stewardship.• Different Cultures: Collaborative efforts and fairness address diverse cul-tural perspectives in environmental practices.• Institutions: Institutional responsibility shapes transparent and fair envi-ronmental stewardship policies.• Nation: Political accountability ensures equitable and responsible envi-ronmental practices nationally.• Globally: Global collaboration fosters shared responsibility for environ-mental stewardship worldwide.• Over Time: Historical perspectives guide context-aware and equitable en-vironmental stewardship evolution.Social Justice and Equality• Individual: Fair and responsible individual actions contribute to socialjustice and equality.• Different Cultures: Collaborative efforts and cultural sensitivity addressdiverse cultural perspectives in promoting equality.• Institutions: Institutional responsibility shapes transparent and fair socialjustice policies.• Nation: Political accountability ensures equitable and responsible socialjustice on a national level.• Globally: Global collaboration fosters shared responsibility for social jus-tice and equality worldwide.• Over Time: Historical perspectives guide context-aware and equitable so-cial justice evolution.161Healthcare Equity• Individual: Responsible individual health practices contribute to equitablehealthcare.Different Cultures: Collaborative efforts and cultural sensitivity addressdiverse cultural health beliefs.• Institutions: Institutional responsibility shapes transparent and fair health-care policies.• Nation: Political accountability ensures equitable healthcare nationally.• Globally: Global collaboration fosters shared responsibility for healthcareequity worldwide.• Over Time: Historical perspectives guide context-aware and equitablehealthcare evolution.Education Access• Individual: Responsible individual learning practices contribute to equi-table education access.• Different Cultures: Collaborative efforts and fairness address diverse cul-tural perspectives in education.• Institutions: Institutional responsibility shapes transparent and fair edu-cation access policies.• Nation: Political accountability ensures equitable education nationally.• Globally: Global collaboration fosters shared responsibility for educationaccess worldwide.• Over Time: Historical perspectives guide context-aware and equitable ed-ucation access evolution.Economic Inclusivity• Individual: Responsible economic practices contribute to economic inclu-sivity. Different Cultures: Collaborative efforts and cultural sensitivityaddress diverse cultural economic models.• Institutions: Institutional responsibility shapes transparent and fair eco-nomic policies.• Nation: Political accountability ensures economic inclusivity on a nationallevel.• Globally: Global collaboration fosters shared responsibility for economicinclusivity worldwide.162• Over Time: Historical perspectives guide context-aware and equitable eco-nomic inclusivity evolution.Technological Ethics• Individual: Responsible individual technology use contributes to ethicaltechnology practices.• Different Cultures: Collaborative efforts and cultural sensitivity addressdiverse cultural perspectives in technological ethics.• Institutions: Institutional responsibility shapes transparent and fair tech-nological policies.• Nation: Political accountability ensures ethical technology use on a na-tional level.• Globally: Global collaboration fosters shared responsibility for technolog-ical ethics worldwide.• Over Time: Historical perspectives guide context-aware and equitabletechnological ethics evolution.Political Accountability• Individual: Responsible individual civic engagement contributes to polit-ical accountability.• Different Cultures: Collaborative efforts and fairness address diverse cul-tural political perspectives.• Institutions: Institutional responsibility shapes transparent and fair po-litical policies.• Nation: Political accountability ensures accountable political processes ona national level.• Globally: Global collaboration fosters shared responsibility for politicalaccountability worldwide.• Over Time: Historical perspectives guide context-aware and equitable po-litical accountability evolution.Local, Inter-Regional and International Collaboration• Individual: Responsible individual collaboration contributes to successfulglobal collaboration.• Different Cultures: Collaborative efforts and cultural sensitivity addressdiverse cultural perspectives in collaborative initiatives.163• Institutions: Institutional responsibility shapes transparent and fair col-laboration policies.• Nation: Political accountability ensures equitable collaboration nationally.• Globally: Global collaboration fosters shared responsibility for collabora-tion worldwide.• Over Time: Historical perspectives guide context-aware and equitable col-laboration evolution.Interactions Across RowsCultural sensitivity in communal values ensures that diverse cultural perspec-tives are respected, creating an inclusive and fair community engagement. Ad-dressing social justice in environmental practices promotes equitable and respon-sible environmental stewardship, considering the impact on diverse communi-ties. Promoting healthcare equity and education access ensures that diversecommunities have fair and transparent access to essential services, contributingto overall well-being. Ethical technology practices contribute to economic inclu-sivity, ensuring fair and responsible economic policies that consider technologi-cal advancements. Political accountability fosters fair collaboration, promotingtransparent and responsible collaboration in political processes.Interactions Across Columns and RowsResponsible individual actions contribute to positive outcomes across all ethicsareas, creating a foundation for transparency, fairness, responsibility, equity, andcontext-awareness. Considering diverse cultural perspectives in individual ac-tions, collaborative efforts, and institutional responsibilities ensures inclusivity,fairness, and cultural sensitivity. Institutional responsibility influences nationalpolicies, shaping the ethical landscape within a country and contributing toresponsible governance and fair practices. Global collaboration and historicalperspectives guide the evolution of ethical practices over time, fostering a glob-ally inclusive, transparent, fair, responsible, and context-aware approach.Multi-interactions of ethicsThe interactions between multiple columns in the multiscale ethics frameworkare dynamic and interwoven, creating a cohesive ethical environment. Trans-parency and fairness at the individual level contribute to fair institutional poli-cies, while acknowledging diverse cultural perspectives promotes responsiblegovernance on a national level. Global responsibility guides the equitable evo-lution of ethical norms over time, and context-awareness influences fair collabo-ration across local, inter-regional, and international scales. The connections be-tween equity and context-awareness extend to various columns, such as health-care equity influencing fair and context-aware education access. Overall, the164principles of transparency, fairness, responsibility, equity, and context-awarenessintersect, shaping ethical behavior across different scales and fostering a holisticand inclusive ethical framework over time.Table 61: Multi-interactions of ethicsEthics Area Individual DifferentCulturesInstitutions Nation Platforms Globally Over TimeCommunityEngagementand Commu-nal ValuesPersonal com-mitmentRespect for di-verse valuesEstablishmentof guidelinesIntegrationinto policiesEngagementin online andoffline com-munitiesRecognitionand apprecia-tion of diversevaluesPreservationand adap-tation overgenerationsCultural Sen-sitivityCulturalawarenessMutual re-spectImplementationof policiesPromotion asa national as-setCultural in-clusivity indigital spacesCollaborationwith diverseinternationalculturesPreservationand evolutionof culturalidentitiesEnvironmentalStewardshipPersonal re-sponsibilityIntegration ofeco-friendlypracticesImplementationof envi-ronmentalpoliciesNational com-mitment tosustainabilityOnline plat-forms pro-moting eco-consciousactionsGlobal co-operation onsustainabledevelopmentSustainablepractices forfuture genera-tionsSocial Justiceand EqualityAdvocacy forequalityRecognitionof diverseperspectivesImplementationof policiesLegal frame-works forsocial justiceOnline plat-forms foradvocacy andawarenessGlobal collab-oration on hu-man rightsProgressiontoward in-creased socialjusticeHealthcareEquityPersonalhealth respon-sibilityCulturallysensitivehealthcarepracticesEqual accesshealthcaresystemsNationalhealth policiesfor equityDigital plat-forms forhealth infor-mationGlobal collab-orations forhealth equitySustainablehealthcare forfuture genera-tionsEducation Ac-cessPersonal com-mitment tolearningInclusionof diverseeducationalperspectivesEqual accesseducationalsystemsNational com-mitment toaccessibleeducationOnline plat-forms fore-learningGlobal part-nerships foreducationopportunitiesEvolving edu-cation systemsover genera-tionsEconomic In-clusivityEthical eco-nomic prac-ticesIntegration ofdiverse eco-nomic modelsEconomicpolicies forinclusivityNational com-mitment to in-clusive growthOnline plat-forms pro-moting faireconomicpracticesCollaborativeefforts forglobal eco-nomic inclu-sivitySustainableeconomicpractices forfuture genera-tionsTechnologicalEthicsResponsibletechnologyuseConsiderationof diversecultural per-spectivesEthical guide-lines and regu-lationsNationalpolicies forresponsibletechnologyDigital plat-forms forethical techdiscussionsInternationalagreements onethical techuseEthical con-siderations intechnologicaladvancementsPolitical Ac-countabilityActive civicengagementRecognitionof diversepolitical per-spectivesTransparentpolitical sys-temsNational com-mitment to ac-countabilityDigital plat-forms forpolitical dis-courseInternationalcollabo-ration fortransparentgovernanceEvolutionof politicalaccountabilityLocal, Inter-regional andInternationalCollaborationAppreciationfor globalperspectivesIntegration ofdiverse inter-national view-pointsGlobal collab-oration frame-worksNational com-mitment tointernationalcooperationOnline plat-forms forcross-culturalcollaborationCollaborativeefforts onglobal chal-lengesBuilding andsustaininginternationalcollaborations8.5 Ethics at the individual level: users, developers, de-signers, individual investorsCommunity Engagement and Communal ValuesIndividual: Transparency involves openly communicating personal values andcommitments to the community. Fairness ensures equal participation, respon-sibility in upholding communal values, and equity in benefit-sharing. Responsi-bility requires individuals to contribute positively to the community. Context-awareness considers the cultural context and its impact on communal values.Cultural SensitivityIndividual: Transparent appreciation of diverse cultures, ensuring fairness inpersonal interactions. Responsibility involves learning and adapting to cultural165nuances. Equity includes respecting cultural differences. Context-awarenessrequires understanding the cultural context of one’s actions.Environmental StewardshipIndividual: Transparent eco-friendly practices, fair resource consumption, re-sponsible waste management, and equitable access to a clean environment. Re-sponsibility involves personal actions for environmental sustainability. Context-awareness considers local ecosystems.Social Justice and EqualityIndividual: Transparent advocacy for social justice, fair treatment of diverseperspectives, and responsible behavior. Equity requires acknowledging and ad-dressing social disparities. Context-awareness involves understanding the uniquechallenges faced by different groups.Healthcare EquityIndividual: Transparency in personal health choices, fair access to healthcareinformation, responsible health practices, and equitable support for healthcareaccess. Context-awareness considers cultural health beliefs and practices.Education AccessIndividual: Transparent commitment to learning, fairness in educational oppor-tunities, responsible study habits, and equity in access to education. Context-awareness involves understanding diverse learning needs.Economic Inclusivity:Individual: Transparent economic practices, fair business dealings, responsiblefinancial decisions, and equitable opportunities. Context-awareness considerslocal economic contexts.Technological EthicsIndividual: Transparent use of technology, fair consideration of diverse perspec-tives, responsible digital behavior, and equitable access to technological benefits.Context-awareness involves understanding the societal impact of technology.Political Accountability:Individual: Transparent civic engagement, fair consideration of political per-spectives, responsible voting, and equity in political representation. Context-awareness involves understanding the political context of one’s actions.166Local, Inter-regional and International CollaborationIndividual: Transparent appreciation for global perspectives, fair collaboration,responsible global citizenship, and equity in international interactions. Context-awareness involves understanding the cultural and political contexts of collabo-rating regions.8.6 Different Cultures, Different groups, Culture-AwareEthicsCommunity Engagement and Communal ValuesDifferent Cultures:• Transparency: Transparent appreciation for diverse cultural values andpractices, openly sharing and learning about different cultural perspectiveswithin the community.• Fairness: Fair treatment of all cultures within the community, ensuringthat cultural diversity is recognized and respected in communal engage-ments.• Responsibility: Responsible engagement with different cultures, takinginitiative to understand and promote cultural awareness within the com-munity.• Equity: Equitable representation of diverse cultural voices, ensuring thatall cultures have an equal opportunity to contribute to communal values.• Context-Awareness: Being context-aware of the cultural nuances and sen-sitivities, adapting community engagements to respect the diversity ofcultures involved.Cultural SensitivityDifferent Cultures:• Transparency: Transparent communication of cultural awareness, acknowl-edging the richness of various cultures, and expressing openness to culturallearning.• Fairness: Fair treatment of all cultures, avoiding biases or preferences, andappreciating the unique aspects of each cultural identity.• Responsibility: Taking responsibility for cultural education, actively chal-lenging stereotypes, and contributing positively to cultural understanding.• Equity: Equitable consideration of all cultures, valuing diversity, and en-suring that no culture is marginalized or unfairly treated.167• Context-Awareness: Being context-aware of cultural contexts, understand-ing historical and social factors shaping cultures, and adapting behaviorto show respect within specific cultural settings.Environmental StewardshipDifferent Cultures:• Transparency: Transparent communication about the environmental im-pact of diverse cultural practices, fostering awareness and open dialogue.• Fairness: Fair consideration of how different cultures interact with theenvironment, ensuring that environmental practices are just and inclusive.• Responsibility: Shared responsibility for environmental sustainability, ac-knowledging the role of diverse cultures in preserving ecosystems.• Equity: Equitable distribution of environmental responsibilities, recogniz-ing that all cultures should contribute to sustainable practices.• Context-Awareness: Being context-aware of local ecosystems, adaptingenvironmental stewardship to respect the unique environmental contextsshaped by different cultures.Social Justice and EqualityDifferent Cultures:• Transparency: Transparent advocacy for social justice, openly addressingcultural disparities, and promoting fairness in social issues.• Fairness: Fair consideration of cultural perspectives in social justice mat-ters, ensuring that diverse voices are heard and valued.• Responsibility: Shared responsibility for addressing cultural inequalities,actively working towards equitable social outcomes.• Equity: Equitable treatment of different cultural groups, acknowledgingand rectifying social injustices across cultures.• Context-Awareness: Being context-aware of unique challenges faced bydifferent cultural groups, adapting social justice efforts to address specificcultural contexts.Healthcare EquityDifferent Cultures:• Transparency: Transparent communication about healthcare practices,openly discussing cultural factors that influence health decisions.168• Fairness: Fair consideration of cultural beliefs in healthcare, ensuring thathealthcare practices are culturally sensitive and inclusive.• Responsibility: Taking responsibility for understanding and respectingdiverse healthcare needs, contributing to culturally competent healthcare.• Equity: Equitable access to healthcare for all cultural groups, addressinghealthcare disparities and promoting inclusive health policies.• Context-Awareness: Being context-aware of cultural health beliefs, adapt-ing healthcare approaches to respect and integrate diverse cultural per-spectives.Education AccessDifferent Cultures:• Transparency: Transparent communication about educational opportu-nities, fostering open dialogue about the importance of education acrossdiverse cultures.• Fairness: Fair consideration of diverse cultural perspectives in education,ensuring that educational resources are accessible to all cultural groups.• Responsibility: Taking responsibility for promoting education across cul-tures, actively contributing to inclusive educational environments.• Equity: Equitable access to educational opportunities for all culturalgroups, addressing educational disparities and promoting inclusive edu-cational policies.• Context-Awareness: Being context-aware of diverse learning needs shapedby cultural backgrounds, adapting educational approaches to respect andaccommodate cultural diversity.Economic InclusivityDifferent Cultures:• Transparency: Transparent communication about economic opportuni-ties, fostering an open discussion about economic practices that respectdiverse cultures.• Fairness: Fair consideration of diverse cultural economic models, ensuringthat economic opportunities are distributed without bias.• Responsibility: Taking responsibility for ethical economic practices acrosscultures, actively contributing to fair economic behavior.• Equity: Equitable economic opportunities for all cultural groups, address-ing economic disparities and promoting inclusive economic policies.169• Context-Awareness: Being context-aware of local economic contexts, adapt-ing economic practices to respect and integrate diverse cultural perspec-tives.Technological EthicsDifferent Cultures:• Transparency: Transparent communication about technological advance-ments, fostering open discussions about how technology impacts diversecultures.• Fairness: Fair consideration of diverse cultural perspectives in technolog-ical development, ensuring that technology is designed without bias.• Responsibility: Taking responsibility for ethical technology use across cul-tures, actively contributing to the responsible development and use oftechnology.• Equity: Equitable access to technological benefits for all cultural groups,addressing digital disparities and promoting inclusive technology policies.• Context-Awareness: Being context-aware of the societal impact of tech-nology on different cultures, adapting technological approaches to respectand accommodate diverse cultural perspectives.Political AccountabilityDifferent Cultures:• Transparency: Transparent communication about political processes, fos-tering open dialogue about the impact of politics on diverse cultures.• Fairness: Fair consideration of diverse political perspectives, ensuring thatpolitical systems are just and inclusive.• Responsibility: Taking responsibility for political engagement across cul-tures, actively contributing to transparent and accountable political pro-cesses.• Equity: Equitable political representation for all cultural groups, address-ing political disparities and promoting inclusive political policies.• Context-Awareness: Being context-aware of the political contexts of dif-ferent cultures, adapting political engagement to respect and integratediverse cultural perspectives.170International Collaboration on MI Ethics in AfricaDifferent Cultures:• Transparency: Transparent communication about collaborative efforts onMI ethics, fostering open dialogue about the impact of MI policies ondiverse cultures within Africa.• Fairness: Fair consideration of diverse cultural perspectives in interna-tional collaborations, ensuring that MI ethics frameworks are just andinclusive across the continent.• Responsibility: Taking responsibility for collaborative MI ethics initia-tives across cultures, actively contributing to transparent and accountableethical frameworks for MI deployment.• Equity: Equitable representation in international collaborations for all cul-tural groups within Africa, addressing disparities and promoting inclusiveMI ethics policies.• Context-Awareness: Being context-aware of the cultural diversity withinAfrica in MI ethics, adapting collaborative efforts to respect and integratediverse cultural perspectives in shaping ethical guidelines for MI.8.7 InstitutionsCommunity Engagement and Communal ValuesInstitutions:• Transparency: Transparent establishment of guidelines for environmentalstewardship initiatives, openly communicating institutional strategies forsustainability within the community.• Fairness: Fair implementation of environmental policies, ensuring that allcommunity members have equal access to information and opportunitiesfor involvement.• Responsibility: Institutional responsibility for fostering positive environ-mental practices within the community, actively engaging in initiatives forsustainable development.• Equity: Equitable distribution of environmental benefits and responsibil-ities among community members, addressing environmental disparities.• Context-Awareness: Instituting context-aware environmental programsthat consider local ecosystems, adapting initiatives to respect the uniqueenvironmental context.171Cultural SensitivityInstitutions:• Transparency: Transparent communication about how institutional prac-tices consider and respect diverse cultural environmental perspectives.• Fairness: Fair integration of cultural values into institutional environmen-tal policies, ensuring that no culture is disproportionately affected.• Responsibility: Institutional responsibility for fostering cultural sensitivityin environmental initiatives, actively addressing cultural implications ofenvironmental decisions.• Equity: Equitable distribution of environmental benefits and burdens,recognizing and rectifying any cultural biases in institutional practices.• Context-Awareness: Developing institution-wide environmental strategiesthat respect and integrate the diverse cultural contexts shaping environ-mental perspectives.Environmental StewardshipInstitutions:• Transparency: Transparent communication of institutional efforts for en-vironmental stewardship, openly sharing goals, progress, and outcomeswith stakeholders.• Fairness: Fair allocation of resources and opportunities for environmentalinitiatives, ensuring that all stakeholders have a fair chance to participate.• Responsibility: Institutional responsibility for implementing effective en-vironmental stewardship practices, actively contributing to global sustain-ability goals.• Equity: Equitable distribution of environmental benefits, addressing anydisparities in the impact of institutional environmental practices.• Context-Awareness: Developing institution-wide environmental strategiesthat consider the unique contexts of local ecosystems and global environ-mental challenges.Social Justice and EqualityInstitutions:• Transparency: Transparent communication of institutional efforts for so-cial justice in the context of environmental initiatives, fostering open dia-logue about the social implications.172• Fairness: Fair consideration of diverse social perspectives in institutionalenvironmental policies, ensuring that social justice is central to sustain-ability practices.• Responsibility: Institutional responsibility for addressing social inequal-ities exacerbated by environmental issues, actively working toward equi-table outcomes.• Equity: Equitable distribution of social and environmental benefits, rec-ognizing and rectifying any social disparities resulting from institutionalpractices.• Context-Awareness: Developing institution-wide environmental strategiesthat are context-aware of unique social challenges, adapting initiatives topromote social justice.Healthcare EquityInstitutions:• Transparency: Transparent communication about how institutional health-care practices consider and respect diverse cultural health beliefs.• Fairness: Fair integration of cultural values into institutional healthcarepolicies, ensuring that healthcare services are culturally sensitive and in-clusive.• Responsibility: Institutional responsibility for fostering healthcare equity,actively addressing cultural implications of healthcare decisions and prac-tices.• Equity: Equitable distribution of healthcare benefits, recognizing and rec-tifying any cultural biases in institutional healthcare services.• Context-Awareness: Developing institution-wide healthcare strategies thatconsider the unique cultural contexts shaping health beliefs and practices.Education AccessInstitutions:• Transparency: Transparent communication of institutional efforts for pro-moting education access, openly sharing strategies and progress with stake-holders.• Fairness: Fair allocation of resources and opportunities for education ac-cess initiatives, ensuring that all stakeholders have a fair chance to par-ticipate.173• Responsibility: Institutional responsibility for implementing effective ed-ucation access practices, actively contributing to inclusive educational en-vironments.• Equity: Equitable distribution of educational benefits, addressing any dis-parities in the impact of institutional education access practices.• Context-Awareness: Developing institution-wide education access strate-gies that are context-aware of diverse learning needs and cultural back-grounds.Economic InclusivityInstitutions:• Transparency: Transparent communication about how institutional eco-nomic practices consider and respect diverse cultural economic models.• Fairness: Fair integration of diverse economic models into institutionaleconomic policies, ensuring that economic opportunities are distributedwithout bias.• Responsibility: Institutional responsibility for fostering economic inclu-sivity, actively addressing cultural implications of economic decisions andpractices.• Equity: Equitable distribution of economic benefits, recognizing and rec-tifying any cultural biases in institutional economic practices.• Context-Awareness: Developing institution-wide economic inclusivity strate-gies that are context-aware of local economic contexts and diverse culturalperspectives.Technological EthicsInstitutions:• Transparency: Transparent communication about institutional technolog-ical advancements, fostering open dialogue about how technology impactsdiverse cultures.• Fairness: Fair consideration of diverse cultural perspectives in institutionaltechnological development, ensuring that technology is designed withoutbias and aligns with ethical standards.• Responsibility: Institutional responsibility for ethical technology use, ac-tively contributing to responsible development and deployment of tech-nology within diverse cultural contexts.• Equity: Equitable access to technological benefits for all cultural groups,addressing digital disparities and promoting inclusive technology policies.174• Context-Awareness: Developing institution-wide technological strategiesthat are context-aware of the societal impact on different cultures, adapt-ing technological approaches to respect and accommodate diverse culturalperspectives.Political AccountabilityInstitutions:• Transparency: Transparent communication about institutional politicalprocesses, fostering open dialogue about the impact of politics on diversecultures.• Fairness: Fair consideration of diverse political perspectives in institu-tional policies, ensuring that political systems are just and inclusive.• Responsibility: Institutional responsibility for political accountability, ac-tively contributing to transparent and accountable political processes withindiverse cultural contexts.• Equity: Equitable political representation for all cultural groups, address-ing political disparities and promoting inclusive political policies.• Context-Awareness: Developing institution-wide political accountabilitystrategies that are context-aware of the political contexts of different cul-tures, adapting political engagement to respect and integrate diverse cul-tural perspectives.Local, Inter-regional and International CollaborationInstitutions:• Transparency: Transparent communication in collaborative efforts, foster-ing open dialogue about the goals, strategies, and outcomes of institutionalcollaborations across cultures.• Fairness: Fair consideration of diverse cultural perspectives in collabora-tive initiatives, ensuring that all collaborators have an equal voice androle.• Responsibility: Institutional responsibility for promoting collaborative ef-forts that respect cultural diversity, actively contributing to positive globalcollaborations.• Equity: Equitable distribution of benefits and responsibilities in collabo-rative initiatives, recognizing and rectifying any cultural biases in institu-tional collaborations.• Context-Awareness: Developing institution-wide collaboration strategiesthat are context-aware of the cultural nuances and challenges in local,inter-regional, and international contexts.1758.8 Ethics at a Nation LevelCommunity Engagement and Communal ValuesNation:• Transparency: The nation promotes transparent communication aboutcommunal values and engagement strategies, ensuring that citizens areinformed about the goals and outcomes of community initiatives.• Fairness: The nation works towards fair distribution of resources and op-portunities within communities, striving to eliminate disparities and pro-mote inclusivity in communal values.• Responsibility: The nation holds a responsibility to support and facilitatepositive communal values, actively engaging in initiatives that contributeto the well-being of communities.• Equity: The nation aims for equitable development across communities,addressing regional disparities and promoting equal access to communalbenefits.• Context-Awareness: The nation considers the unique cultural and geo-graphical contexts of different communities, adapting national strategiesto respect and accommodate diversity.Cultural SensitivityNation:• Transparency: The nation transparently communicates its commitmentto cultural sensitivity, ensuring that policies and initiatives consider andrespect diverse cultural perspectives.• Fairness: The nation strives for fair treatment of all cultures, fostering anenvironment where cultural diversity is celebrated and valued.• Responsibility: The nation takes responsibility for fostering a culturallysensitive environment, actively working to eliminate cultural biases in na-tional policies.• Equity: The nation aims for equitable representation of cultural voices innational decision-making, recognizing and addressing any cultural dispar-ities.• Context-Awareness: The nation develops cultural policies that are context-aware, considering the historical and social context of different culturalgroups within the nation.176Environmental StewardshipNation:• Transparency: The nation communicates transparently about its environ-mental stewardship efforts, sharing information about national strategiesfor sustainability.• Fairness: The nation ensures fair distribution of environmental benefitsand responsibilities, working to eliminate environmental injustices acrossregions.• Responsibility: The nation takes responsibility for implementing effectivenational environmental stewardship practices, contributing to global sus-tainability goals.• Equity: The nation aims for equitable environmental policies, addressingany disparities in the impact of national environmental practices.• Context-Awareness: The nation develops national environmental strate-gies that consider the unique contexts of local ecosystems and diversegeographical regions.Social Justice and EqualityNation:• Transparency: The nation communicates transparently about its effortsfor social justice, fostering open dialogue about the social implications ofnational policies.• Fairness: The nation ensures fair consideration of diverse social perspec-tives in national policies, striving for just and inclusive outcomes.• Responsibility: The nation takes responsibility for addressing social in-equalities exacerbated by national issues, actively working toward equi-table social outcomes.• Equity: The nation aims for equitable distribution of social benefits, rec-ognizing and rectifying any social disparities resulting from national prac-tices.• Context-Awareness: The nation develops national strategies that are context-aware of unique social challenges, adapting policies to promote social jus-tice.177Healthcare EquityNation:• Transparency: The nation communicates transparently about how na-tional healthcare practices consider and respect diverse cultural healthbeliefs.• Fairness: The nation ensures fair integration of cultural values into na-tional healthcare policies, promoting healthcare services that are culturallysensitive and inclusive.• Responsibility: The nation takes responsibility for fostering healthcareequity, actively addressing cultural implications of national healthcare de-cisions and practices.• Equity: The nation aims for equitable distribution of healthcare bene-fits, recognizing and rectifying any cultural biases in national healthcareservices.• Context-Awareness: The nation develops national healthcare strategiesthat are context-aware, considering the unique cultural contexts shapinghealth beliefs and practices.Education AccessNation:• Transparency: The nation transparently communicates its efforts to pro-mote education access, ensuring that citizens are informed about nationalstrategies and progress.• Fairness: The nation ensures fair allocation of resources and opportuni-ties for education access initiatives, striving to eliminate disparities andpromote inclusivity in education.• Responsibility: The nation takes responsibility for implementing effectivenational education access practices, actively contributing to inclusive ed-ucational environments.• Equity: The nation aims for equitable distribution of educational bene-fits, addressing any disparities in the impact of national education accesspractices.• Context-Awareness: The nation develops national education access strate-gies that are context-aware, considering diverse learning needs shaped bycultural backgrounds.178Economic InclusivityNation:• Transparency: The nation transparently communicates about how na-tional economic practices consider and respect diverse cultural economicmodels.• Fairness: The nation ensures fair integration of diverse economic modelsinto national economic policies, promoting economic opportunities thatare distributed without bias.• Responsibility: The nation takes responsibility for fostering economic in-clusivity, actively addressing cultural implications of national economicdecisions and practices.• Equity: The nation aims for equitable distribution of economic benefits,recognizing and rectifying any cultural biases in national economic prac-tices.• Context-Awareness: The nation develops national economic inclusivitystrategies that are context-aware, considering local economic contexts anddiverse cultural perspectives.Technological EthicsNation:• Transparency: The nation transparently communicates about nationaltechnological advancements, fostering open dialogue about how technologyimpacts diverse cultures.• Fairness: The nation ensures fair consideration of diverse cultural perspec-tives in national technological development, striving for technology thatis designed without bias.• Responsibility: The nation takes responsibility for ethical technology use,actively contributing to the responsible development and deployment oftechnology within diverse cultural contexts.• Equity: The nation aims for equitable access to technological benefits forall cultural groups, addressing digital disparities and promoting inclusivetechnology policies.• Context-Awareness: The nation develops national technological strategiesthat are context-aware, considering the societal impact of technology ondifferent cultures and adapting technological approaches to respect andaccommodate diverse cultural perspectives.179Political AccountabilityNation:• Transparency: The nation communicates transparently about nationalpolitical processes, fostering open dialogue about the impact of politics ondiverse cultures.• Fairness: The nation ensures fair consideration of diverse political per-spectives in national policies, striving for political systems that are justand inclusive.• Responsibility: The nation takes responsibility for political accountability,actively contributing to transparent and accountable political processeswithin diverse cultural contexts.• Equity: The nation aims for equitable political representation for all cul-tural groups, addressing political disparities and promoting inclusive po-litical policies.• Context-Awareness: The nation develops national political accountabilitystrategies that are context-aware, considering the political contexts of dif-ferent cultures and adapting political engagement to respect and integratediverse cultural perspectives.Local, Inter-regional and International CollaborationNation:• Transparency: The nation transparently communicates in collaborativeefforts, fostering open dialogue about the goals, strategies, and outcomesof national collaborations across cultures.• Fairness: The nation ensures fair consideration of diverse cultural perspec-tives in national collaborative initiatives, striving for equal representationand contributions.• Responsibility: The nation takes responsibility for promoting collabora-tive efforts that respect cultural diversity, actively contributing to positiveglobal collaborations.• Equity: The nation aims for equitable distribution of benefits and respon-sibilities in national collaborative initiatives, recognizing and rectifyingany cultural biases.• Context-Awareness: The nation develops national collaboration strategiesthat are context-aware, considering the cultural nuances and challenges inlocal, inter-regional, and international contexts.1808.9 Ethics at the Platforms levelCommunity Engagement and Communal ValuesOn and off-line platforms:• Transparency: On and off-line platforms facilitate transparent commu-nication about community engagement, ensuring that information aboutcommunal values is accessible and openly discussed.• Fairness: Platforms strive to provide fair representation of diverse voices,fostering an inclusive space where different perspectives on communal val-ues are acknowledged and respected.• Responsibility: Platforms take responsibility for maintaining a positivecommunity engagement environment, actively moderating discussions andaddressing any issues that may arise.• Equity: Platforms work towards equitable participation, ensuring thatindividuals from all backgrounds have equal opportunities to engage andcontribute to communal values discussions.• Context-Awareness: Platforms are context-aware, adapting to the cul-tural nuances and diversity of perspectives within community engagementdiscussions both online and offline.Cultural SensitivityOn and off-line platforms:• Transparency: Platforms transparently communicate their commitmentto cultural sensitivity, providing clear guidelines on respectful interactionsand content sharing that considers diverse cultural perspectives.• Fairness: Platforms aim for fair representation and treatment of diversecultural content, fostering an environment where cultural sensitivity is acentral aspect of online and offline discussions.• Responsibility: Platforms take responsibility for promoting cultural sen-sitivity, actively moderating content to eliminate cultural biases and en-suring respectful interactions among users.• Equity: Platforms work towards equitable access to cultural content, ad-dressing any disparities and promoting the inclusion of various culturalperspectives in online and offline spaces.• Context-Awareness: Platforms are context-aware, adapting their guide-lines and content moderation practices to respect the unique cultural con-texts shaping online and offline discussions.181Environmental StewardshipOn and off-line platforms:• Transparency: Platforms transparently share information about environ-mental stewardship efforts, utilizing online spaces to disseminate knowl-edge and updates on national and global sustainability initiatives.• Fairness: Platforms promote fair and unbiased coverage of environmentalissues, ensuring that discussions on environmental stewardship representdiverse perspectives and solutions.• Responsibility: Platforms take responsibility for encouraging positive envi-ronmental practices, actively engaging users in discussions and initiativesthat contribute to environmental sustainability.• Equity: Platforms aim for equitable distribution of environmental infor-mation, addressing any disparities in access to discussions and resourcesrelated to environmental stewardship.• Context-Awareness: Platforms are context-aware, adapting their approachto environmental discussions to consider the global and local contextsshaping sustainability efforts.Social Justice and EqualityOn and off-line platforms:• Transparency: Platforms transparently communicate efforts for social jus-tice, utilizing online spaces to share information about initiatives and ad-vocate for equality.• Fairness: Platforms work towards fair representation of diverse social per-spectives, fostering an inclusive environment where discussions on socialjustice are free from biases.• Responsibility: Platforms take responsibility for addressing social inequal-ities, actively moderating content and discussions to promote a just andequal online and offline space.• Equity: Platforms aim for equitable access to social justice discussions,recognizing and rectifying any disparities in the impact of online and offlinepractices.• Context-Awareness: Platforms are context-aware, adapting their approachto social justice discussions to consider the unique social challenges andcontexts influencing equality.182Healthcare EquityOn and off-line platforms:• Transparency: Platforms transparently communicate about how health-care practices consider and respect diverse cultural health beliefs, dissem-inating information about inclusive healthcare initiatives.• Fairness: Platforms ensure fair integration of cultural values into health-care discussions, fostering an online environment where healthcare servicesare culturally sensitive and inclusive.• Responsibility: Platforms take responsibility for promoting healthcare eq-uity, actively engaging users in discussions and initiatives that addresscultural implications of healthcare decisions and practices.• Equity: Platforms aim for equitable access to healthcare discussions andinformation, recognizing and rectifying any cultural biases in online andoffline healthcare spaces.• Context-Awareness: Platforms are context-aware, adapting their health-care discussions to consider the unique cultural contexts shaping healthbeliefs and practices in online and offline settings.Education AccessOn and off-line platforms:• Transparency: Platforms transparently communicate efforts to promoteeducation access, utilizing online spaces to share information about initia-tives and progress in making education accessible.• Fairness: Platforms ensure fair allocation of resources and opportunitiesfor education access initiatives, fostering an online environment where allusers have equal chances to participate in educational discussions.• Responsibility: Platforms take responsibility for implementing effectiveeducation access practices, actively contributing to inclusive educationalenvironments both online and offline.• Equity: Platforms aim for equitable distribution of educational benefits,addressing any disparities in the impact of online and offline educationaccess practices.• Context-Awareness: Platforms are context-aware, adapting their approachto education access discussions to consider diverse learning needs and cul-tural backgrounds in online and offline spaces.183Economic InclusivityOn and off-line platforms:• Transparency: Platforms transparently communicate about how economicpractices consider and respect diverse cultural economic models, dissemi-nating information about inclusive economic initiatives.• Fairness: Platforms ensure fair integration of diverse economic modelsinto economic discussions, fostering an online environment where economicopportunities are distributed without bias.• Responsibility: Platforms take responsibility for promoting economic in-clusivity, actively engaging users in discussions and initiatives that addresscultural implications of economic decisions and practices.• Equity: Platforms aim for equitable access to economic discussions andbenefits, recognizing and rectifying any cultural biases in online and of-fline economic spaces. Context-Awareness: Platforms are context-aware,adapting their economic discussions to consider local economic contextsand diverse cultural perspectives in online and offline settings.Technological EthicsOn and off-line platforms:• Transparency: Platforms transparently communicate about national tech-nological advancements, fostering open dialogue about how technologyimpacts diverse cultures in online and offline spaces.• Fairness: Platforms ensure fair consideration of diverse cultural perspec-tives in technological discussions, striving for technology that is designedwithout bias in both online and offline environments.• Responsibility: Platforms take responsibility for ethical technology use,actively contributing to the responsible development and deployment oftechnology within diverse cultural contexts online and offline.• Equity: Platforms aim for equitable access to technological benefits forall cultural groups, addressing digital disparities and promoting inclusivetechnology policies in both online and offline spaces.• Context-Awareness: Platforms are context-aware, adapting their techno-logical discussions to consider the societal impact of technology on differentcultures in both online and offline settings.184Political AccountabilityOn and off-line platforms:• Transparency: Platforms transparently communicate about national po-litical processes, fostering open dialogue about the impact of politics ondiverse cultures in online and offline spaces.• Fairness: Platforms ensure fair consideration of diverse political perspec-tives in political discussions, striving for just and inclusive political sys-tems both online and offline.• Responsibility: Platforms take responsibility for political accountability,actively contributing to transparent and accountable political processeswithin diverse cultural contexts online and offline.• Equity: Platforms aim for equitable political representation for all cul-tural groups in political discussions, addressing political disparities andpromoting inclusive political policies both online and offline.• Context-Awareness: Platforms are context-aware, adapting their politi-cal discussions to consider the political contexts of different cultures andadapting political engagement to respect and integrate diverse culturalperspectives in both online and offline spaces.Local, Inter-regional and International CollaborationOn and off-line platforms:• Transparency: Platforms transparently communicate in collaborative ef-forts, fostering open dialogue about the goals, strategies, and outcomes ofcollaborative initiatives across cultures in both online and offline spaces.• Fairness: Platforms ensure fair consideration of diverse cultural perspec-tives in collaborative initiatives, striving for equal representation and con-tributions both online and offline.• Responsibility: Platforms take responsibility for promoting collaborativeefforts that respect cultural diversity, actively contributing to positiveglobal collaborations in both online and offline spaces.• Equity: Platforms aim for equitable distribution of benefits and responsi-bilities in collaborative initiatives, recognizing and rectifying any culturalbiases both online and offline.• Context-Awareness: Platforms are context-aware, adapting their collabo-ration strategies to consider the cultural nuances and challenges in local,inter-regional, and international contexts both online and offline.1858.10 What ethics at the globe level?Think global strategies, Act locally ?Community Engagement and Communal ValuesGlobally:• Transparency: Global engagement platforms ensure transparency by fa-cilitating open communication about communal values on a global scale,fostering a cross-cultural exchange of ideas and perspectives.• Fairness: Platforms strive for fairness by promoting equal representationof diverse global voices, creating an inclusive space where different culturescontribute to the discussion on communal values.• Responsibility: Global platforms take responsibility for fostering positiveglobal communal values, actively moderating discussions to ensure respect-ful cross-cultural interactions and addressing any global issues that mayarise.• Equity: Platforms work towards equitable global participation, ensuringthat individuals from all corners of the world have equal opportunities toengage and contribute to discussions on communal values.• Context-Awareness: Global platforms are context-aware, adapting to thecultural nuances and diversity of perspectives within global communityengagement discussions.Cultural SensitivityGlobally:• Transparency: Global platforms transparently communicate their com-mitment to cultural sensitivity, providing guidelines on respectful globalinteractions and content sharing that considers diverse cultural perspec-tives.• Fairness: Platforms aim for fair representation and treatment of diverseglobal cultural content, fostering a worldwide environment where culturalsensitivity is a central aspect of online discussions.• Responsibility: Global platforms take responsibility for promoting cul-tural sensitivity globally, actively moderating content to eliminate culturalbiases and ensuring respectful interactions among users from different cul-tures.• Equity: Platforms work towards equitable access to global cultural con-tent, addressing any disparities and promoting the inclusion of variousglobal cultural perspectives in online spaces.186• Context-Awareness: Global platforms are context-aware, adapting theirguidelines and content moderation practices to respect the unique culturalcontexts shaping global online discussions.Environmental StewardshipGlobally:• Transparency: Global platforms transparently share information aboutenvironmental stewardship efforts, utilizing online spaces to disseminateknowledge and updates on international sustainability initiatives.• Fairness: Platforms promote fair and unbiased coverage of global envi-ronmental issues, ensuring that discussions on environmental stewardshiprepresent diverse global perspectives and solutions.• Responsibility: Platforms take responsibility for encouraging positive globalenvironmental practices, actively engaging users in discussions and initia-tives that contribute to international environmental sustainability.• Equity: Platforms aim for equitable distribution of global environmen-tal information, addressing any disparities in access to discussions andresources related to international environmental stewardship.• Context-Awareness: Global platforms are context-aware, adapting theirapproach to global environmental discussions to consider the global andlocal contexts shaping sustainability efforts.Social Justice and EqualityGlobally:• Transparency: Platforms transparently communicate efforts for global so-cial justice, utilizing online spaces to share information about internationalinitiatives and advocate for equality.• Fairness: Platforms work towards fair representation of diverse global so-cial perspectives, fostering an inclusive environment where discussions onglobal social justice are free from biases.• Responsibility: Platforms take responsibility for addressing global socialinequalities, actively moderating content and discussions to promote a justand equal global online space.• Equity: Platforms aim for equitable access to global social justice dis-cussions, recognizing and rectifying any global disparities resulting fromonline and offline practices.• Context-Awareness: Platforms are context-aware, adapting their approachto global social justice discussions to consider the unique global social chal-lenges and contexts influencing equality.187Healthcare EquityGlobally:• Transparency: Platforms transparently communicate about how globalhealthcare practices consider and respect diverse cultural health beliefs,disseminating information about inclusive healthcare initiatives world-wide.• Fairness: Platforms ensure fair integration of cultural values into globalhealthcare discussions, fostering an online environment where healthcareservices are culturally sensitive and inclusive on a global scale.• Responsibility: Platforms take responsibility for promoting global health-care equity, actively engaging users in discussions and initiatives that ad-dress cultural implications of global healthcare decisions and practices.• Equity: Platforms aim for equitable access to global healthcare discussionsand information, recognizing and rectifying any cultural biases in onlineand offline global healthcare spaces.• Context-Awareness: Platforms are context-aware, adapting their health-care discussions to consider the unique global cultural contexts shapinghealth beliefs and practices.Education AccessGlobally:• Transparency: Platforms transparently communicate efforts to promoteglobal education access, utilizing online spaces to share information aboutinternational initiatives and progress in making education accessible world-wide.• Fairness: Platforms ensure fair allocation of resources and opportunitiesfor global education access initiatives, fostering an online environmentwhere all users have equal chances to participate in global educationaldiscussions.• Responsibility: Platforms take responsibility for implementing effectiveglobal education access practices, actively contributing to inclusive edu-cational environments both online and offline on a global scale.• Equity: Platforms aim for equitable distribution of global educationalbenefits, addressing any global disparities in the impact of online andoffline global education access practices.• Context-Awareness: Platforms are context-aware, adapting their approachto global education access discussions to consider diverse learning needsand cultural backgrounds on a global scale.188Economic InclusivityGlobally:• Transparency: Platforms transparently communicate about how globaleconomic practices consider and respect diverse global economic models,disseminating information about inclusive economic initiatives worldwide.Fairness: Platforms ensure fair integration of diverse global economic mod-els into economic discussions, fostering an online environment where eco-nomic opportunities are distributed without bias on a global scale.• Responsibility: Platforms take responsibility for promoting global eco-nomic inclusivity, actively engaging users in discussions and initiativesthat address cultural implications of global economic decisions and prac-tices.• Equity: Platforms aim for equitable access to global economic discussionsand benefits, recognizing and rectifying any cultural biases in online andoffline global economic spaces.• Context-Awareness: Platforms are context-aware, adapting their economicdiscussions to consider local economic contexts and diverse cultural per-spectives in online and offline global settings.Technological EthicsGlobally:• Transparency: Platforms transparently communicate about global tech-nological advancements, fostering open dialogue about how technologyimpacts diverse cultures in online and offline spaces worldwide.• Fairness: Platforms ensure fair consideration of diverse global culturalperspectives in technological discussions, striving for technology that isdesigned without bias on a global scale.• Responsibility: Platforms take responsibility for ethical technology useglobally, actively contributing to the responsible development and deploy-ment of technology within diverse cultural contexts online and offline.• Equity: Platforms aim for equitable access to global technological ben-efits for all cultural groups, addressing digital disparities and promotinginclusive technology policies in both online and offline global spaces.• Context-Awareness: Platforms are context-aware, adapting their techno-logical discussions to consider the societal impact of technology on differentcultures in both online and offline global settings.189Political AccountabilityGlobally:• Transparency: Platforms transparently communicate about global polit-ical processes, fostering open dialogue about the impact of politics ondiverse cultures in online and offline spaces worldwide.• Fairness: Platforms ensure fair consideration of diverse global politicalperspectives in political discussions, striving for just and inclusive politicalsystems both online and offline.• Responsibility: Platforms take responsibility for political accountabilityglobally, actively contributing to transparent and accountable politicalprocesses within diverse cultural contexts online and offline.• Equity: Platforms aim for equitable political representation for all culturalgroups in global political discussions, addressing political disparities andpromoting inclusive political policies both online and offline.• Context-Awareness: Platforms are context-aware, adapting their politicaldiscussions to consider the political contexts of different cultures globallyand adapting political engagement to respect and integrate diverse culturalperspectives.Local, Inter-regional and International CollaborationGlobally:• Transparency: Platforms transparently communicate in collaborative ef-forts on a global scale, fostering open dialogue about the goals, strategies,and outcomes of collaborative initiatives across cultures in both onlineand offline spaces.• Fairness: Platforms ensure fair consideration of diverse cultural perspec-tives in collaborative initiatives, striving for equal representation and con-tributions on a global scale both online and offline.• Responsibility: Platforms take responsibility for promoting collaborativeefforts that respect cultural diversity globally, actively contributing topositive global collaborations both online and offline.• Equity: Platforms aim for equitable distribution of benefits and responsi-bilities in collaborative initiatives globally, recognizing and rectifying anycultural biases both online and offline.• Context-Awareness: Platforms are context-aware, adapting their collabo-ration strategies to consider the cultural nuances and challenges in local,inter-regional, and international contexts on a global scale both online andoffline.1908.11 Ethics over time, intergenerational ethicsCommunity Engagement and Communal ValuesOver Time (Intergenerational):• Transparency: Over time, platforms facilitate transparent communicationabout communal values, ensuring that information about historical per-spectives on communal values is accessible and openly discussed.• Fairness: Platforms strive for fairness by preserving historical communalvalues and providing an inclusive space where different historical perspec-tives contribute to the ongoing discussion.• Responsibility: Platforms take responsibility for maintaining a positivehistorical community engagement environment, actively preserving andmoderating discussions to address any issues that may arise over time.• Equity: Platforms work towards equitable intergenerational participation,ensuring that historical voices and perspectives have a place in the ongoingdialogue about communal values.• Context-Awareness: Platforms are context-aware over time, adapting tothe changing cultural contexts and ensuring that historical perspectivesare respected within ongoing community engagement discussions.Cultural SensitivityOver Time (Intergenerational):• Transparency: Platforms transparently communicate their commitmentto cultural sensitivity over time, providing guidelines on respectful inter-actions and content sharing that considers evolving cultural perspectives.• Fairness: Platforms aim for fair representation and treatment of diversecultural content over time, fostering an environment where cultural sensi-tivity is a continuous aspect of discussions across generations.• Responsibility: Platforms take responsibility for promoting cultural sensi-tivity over time, actively preserving historical cultural content and ensur-ing respectful interactions among users from different historical culturalcontexts.• Equity: Platforms work towards equitable access to cultural content overtime, addressing any disparities and promoting the inclusion of varioushistorical cultural perspectives in discussions.• Context-Awareness: Platforms are context-aware over time, adapting theirguidelines and content moderation practices to respect the changing cul-tural nuances and diversity of perspectives within intergenerational dis-cussions.191Environmental StewardshipOver Time (Intergenerational):• Transparency: Platforms transparently share information about the evolu-tion of environmental stewardship efforts over time, utilizing online spacesto disseminate knowledge and updates on the historical progression of sus-tainability initiatives.• Fairness: Platforms promote fair and unbiased coverage of historical envi-ronmental issues, ensuring that discussions on environmental stewardshiprepresent diverse historical perspectives and solutions.• Responsibility: Platforms take responsibility for encouraging positive his-torical environmental practices, actively engaging users in discussions andinitiatives that contribute to the intergenerational sustainability of theplanet.• Equity: Platforms aim for equitable distribution of historical environmen-tal information, addressing any disparities in access to discussions andresources related to historical environmental stewardship.• Context-Awareness: Platforms are context-aware over time, adapting theirapproach to historical environmental discussions to consider the global andlocal historical contexts shaping sustainability efforts.Social Justice and EqualityOver Time (Intergenerational):• Transparency: Platforms transparently communicate efforts for historicalglobal social justice, utilizing online spaces to share information aboutpast initiatives and advocacy for equality across generations.• Fairness: Platforms work towards fair representation of diverse histori-cal global social perspectives, fostering an inclusive environment wherediscussions on historical global social justice are free from biases.• Responsibility: Platforms take responsibility for addressing historical globalsocial inequalities, actively moderating content and discussions to promotea just and equal online space over time.• Equity: Platforms aim for equitable access to historical global social jus-tice discussions, recognizing and rectifying any historical global disparitiesresulting from online and offline practices.• Context-Awareness: Platforms are context-aware over time, adapting theirapproach to historical global social justice discussions to consider theunique historical global social challenges and contexts influencing equality.192Healthcare EquityOver Time (Intergenerational):• Transparency: Platforms transparently communicate about how health-care practices consider and respect diverse cultural health beliefs over time,disseminating information about inclusive healthcare initiatives across gen-erations.• Fairness: Platforms ensure fair integration of cultural values into historicalhealthcare discussions, fostering an online environment where healthcareservices are culturally sensitive and inclusive over time.• Responsibility: Platforms take responsibility for promoting intergenera-tional healthcare equity, actively engaging users in discussions and initia-tives that address cultural implications of historical healthcare decisionsand practices.• Equity: Platforms aim for equitable access to historical healthcare discus-sions and information, recognizing and rectifying any cultural biases inonline and offline historical healthcare spaces.• Context-Awareness: Platforms are context-aware over time, adapting theirhealthcare discussions to consider the unique historical cultural contextsshaping health beliefs and practices.Education AccessOver Time (Intergenerational):• Transparency: Platforms transparently communicate efforts to promotehistorical global education access, utilizing online spaces to share informa-tion about past international initiatives and progress in making educationaccessible worldwide over time.• Fairness: Platforms ensure fair allocation of resources and opportunitiesfor historical global education access initiatives, fostering an online envi-ronment where all users have equal chances to participate in global edu-cational discussions over time.• Responsibility: Platforms take responsibility for implementing effectivehistorical global education access practices, actively contributing to inclu-sive educational environments both online and offline on a global scaleover time.• Equity: Platforms aim for equitable distribution of historical global edu-cational benefits, addressing any global disparities in the impact of onlineand offline global education access practices over time.• Context-Awareness: Platforms are context-aware over time, adapting theirapproach to global education access discussions to consider diverse learn-ing needs and cultural backgrounds on a global scale.193Economic InclusivityOver Time (Intergenerational):• Transparency: Platforms transparently communicate about how globaleconomic practices consider and respect diverse global economic modelsover time, disseminating information about past inclusive economic initia-tives worldwide.• Fairness: Platforms ensure fair integration of diverse global economic mod-els into historical economic discussions, fostering an online environmentwhere economic opportunities are distributed without bias on a globalscale over time.• Responsibility: Platforms take responsibility for promoting historical globaleconomic inclusivity, actively engaging users in discussions and initiativesthat address cultural implications of past global economic decisions andpractices.• Equity: Platforms aim for equitable access to historical global economicdiscussions and benefits, recognizing and rectifying any cultural biases inonline and offline historical global economic spaces.• Context-Awareness: Platforms are context-aware over time, adapting theireconomic discussions to consider local economic contexts and diverse cul-tural perspectives in online and offline global settings.Technological EthicsOver Time (Intergenerational):• Transparency: Platforms transparently communicate about global techno-logical advancements over time, fostering open dialogue about how tech-nology has evolved and impacted diverse cultures in online and offlinespaces worldwide.• Fairness: Platforms ensure fair consideration of diverse global culturalperspectives in historical technological discussions, striving for technologythat is designed without bias on a global scale over time.• Responsibility: Platforms take responsibility for ethical technology useglobally over time, actively contributing to the responsible developmentand deployment of technology within diverse cultural contexts online andoffline.• Equity: Platforms aim for equitable access to historical technological ben-efits for all cultural groups, addressing digital disparities and promotinginclusive technology policies in both online and offline historical globalspaces.194• Context-Awareness: Platforms are context-aware over time, adapting theirtechnological discussions to consider the societal impact of technology ondifferent cultures in both online and offline historical global settings.Political AccountabilityOver Time (Intergenerational):• Transparency: Platforms transparently communicate about global politi-cal processes over time, fostering open dialogue about how political land-scapes have evolved and impacted diverse cultures in online and offlinespaces worldwide.• Fairness: Platforms ensure fair consideration of diverse global politicalperspectives in historical political discussions, striving for just and inclu-sive political systems both online and offline over time.• Responsibility: Platforms take responsibility for political accountabilityglobally over time, actively contributing to transparent and accountablepolitical processes within diverse cultural contexts online and offline.• Equity: Platforms aim for equitable political representation for all culturalgroups in historical global political discussions, addressing political dispar-ities and promoting inclusive political policies both online and offline overtime.• Context-Awareness: Platforms are context-aware over time, adapting theirpolitical discussions to consider the historical political contexts of differ-ent cultures globally and adapting political engagement to respect andintegrate diverse cultural perspectives.Local, Inter-regional and International CollaborationOver Time (Intergenerational):• Transparency: Platforms transparently communicate in collaborative ef-forts on a global scale over time, fostering open dialogue about the goals,strategies, and outcomes of collaborative initiatives across cultures in bothonline and offline spaces.• Fairness: Platforms ensure fair consideration of diverse cultural perspec-tives in collaborative initiatives over time, striving for equal representationand contributions on a global scale both online and offline.• Responsibility: Platforms take responsibility for promoting collaborativeefforts that respect cultural diversity globally over time, actively contribut-ing to positive global collaborations both online and offline.195• Equity: Platforms aim for equitable distribution of benefits and respon-sibilities in collaborative initiatives globally over time, recognizing andrectifying any cultural biases both online and offline.• Context-Awareness: Platforms are context-aware over time, adapting theircollaboration strategies to consider the cultural nuances and challenges inlocal, inter-regional, and international contexts on a global scale both on-line and offline.9 MI in Africa is not only about competitionWhat is a co-opetition ?Co-opetition is a strategy where competitors collaborate on certain projects orinitiatives while simultaneously competing in other areas. It involves a combina-tion of cooperation and competition to achieve mutual benefits for the involvedparties. This strategy recognizes that competitors can sometimes create valueby working together on common goals, even as they compete in other aspects oftheir business. Co-opetition is often seen in industries where there are sharedinterests, such as standard-setting, research and development, or addressingcommon challenges.What is co-opetitive game theory ?Co-opetitive game theory is an extension of traditional game theory that ex-plores scenarios where actors engage in both cooperative and competitive inter-actions simultaneously. In co-opetitive situations, participants collaborate oncertain aspects of a game while competing in others.What is Co-opetitive Mean-Field-Type Game Theory?Co-opetitive mean-field-type game theory explores scenarios where actors en-gage in both cooperative and competitive interactions simultaneously. In co-opetitive situations of mean-field type, participants collaborate on certain as-pects of a game while competing in others. What makes them of mean-fieldtype is their dynamics and/or payoff functions, which depend not only on state-action pairs but also on the distribution of state-action pairs. These games donot necessarily consider a large number of decision-makers. They do not needto assume symmetry, exchangeability, or indistinguishability per type. Thesegames capture more interactions in real-world applications that are risk-awareand non-symmetric, involving both cooperation, competition, partial altruism,partial spite, etc.196Multiscale multimodal Ethics as co-opetitionThe multiscale multimodal Ethics can be seen in Co-opetitive Mean-Field-TypeGame Theory through various dimensions: Individual:• Transparency: Individuals participating in co-opetitive scenarios needtransparency in understanding the collaborative and competitive aspects,fostering trust among participants.• Fairness: Ensuring fairness in the distribution of benefits and competitionis crucial to maintaining a cooperative and competitive balance.• Responsibility: Participants must take responsibility for their actions, con-tributing to both collaborative and competitive elements responsibly.• Equity: Striving for equity in the outcomes of cooperation and competi-tion, considering the diverse interests and capabilities of individuals.Different Cultures:• Transparency: Cultural transparency is essential to bridge understandingamong participants from different cultural backgrounds, creating a sharedunderstanding of co-opetitive dynamics.• Fairness: Ensuring fairness in cultural representation and acknowledgingdiverse perspectives in both collaboration and competition.• Responsibility: Culturally responsible actions contribute to effective co-opetition, respecting the values and norms of diverse cultures.• Equity: Addressing cultural disparities and striving for equitable partici-pation and benefits across diverse cultural groups.Institutions:• Transparency: Transparent institutional processes are necessary to estab-lish clear rules and guidelines for co-opetition.• Fairness: Institutional fairness ensures that rules and policies promoteequitable opportunities for collaboration and competition.• Responsibility: Institutions play a role in enforcing responsible behavior,holding participants accountable for their contributions.• Equity: Institutional structures should strive for equitable distribution ofresources and opportunities within co-opetitive frameworks.Nation:• Transparency: National transparency in co-opetitive initiatives fosterstrust and understanding among participants from different nations.197• Fairness: Ensuring that national interests are fairly represented and con-sidered in co-opetitive dynamics.• Responsibility: National responsibility involves promoting ethical behav-ior and adherence to agreed-upon rules in co-opetition.• Equity: Striving for equitable outcomes that benefit participating nationsand consider their unique circumstances.On and Off-Line Platforms:•• Transparency: Platforms facilitating co-opetition must be transparent intheir operations, ensuring participants understand the online and offlinedynamics.• Fairness: Ensuring fair representation and opportunities for collaborationand competition on digital platforms.• Responsibility: Online platforms should encourage responsible behavior,emphasizing ethical conduct in co-opetitive interactions.• Equity: Addressing digital divides and ensuring equitable access to onlineplatforms for diverse participants.Globally:• Transparency: Global transparency is essential for cross-border co-opetition,ensuring a clear understanding of shared goals and competition.• Fairness: Striving for fairness in global representation and acknowledgingthe diverse interests and contributions of participants worldwide.• Responsibility: Global responsibility involves ethical behavior on a globalscale, promoting positive outcomes in co-opetition.• Equity: Addressing global inequalities and promoting equitable participa-tion and benefits in co-opetitive scenarios.Over Time:• Transparency: Transparent communication over time ensures that partic-ipants are aware of evolving co-opetitive dynamics.• Fairness: Adapting co-opetition strategies to changing circumstances whileensuring fairness in long-term collaborations and competitions.• Responsibility: Long-term responsibility involves sustained ethical behav-ior and commitment to mutual benefits over time.• Equity: Ensuring that co-opetition evolves to address changing needs andmaintains equitable outcomes over time.198MI deployment and adoptionThe deployment of MI in Africa is characterized as not only a competition butalso a coopetitive Mean-Field-Type Game due to several factors:• Cooperative Elements: MI deployment often involves collaboration amongvarious stakeholders, including governments, private enterprises, and re-search institutions. Shared initiatives for technology adoption, skill devel-opment, and addressing common challenges create cooperative dynamics.• Competitive Aspects: Different entities within Africa may compete forresources, market dominance, or technological advancements in the MIlandscape. This competition can drive innovation, economic growth, andthe pursuit of leadership positions in the MI sector.• Mean-Field-Type Dynamics: The mean-field-type dynamics in MI deploy-ment consider not only the actions of individual decision-makers but alsothe overall distribution of state-action pairs. This reflects the intercon-nectedness of MI strategies, where the outcomes depend on collective in-teractions rather than isolated decision-making.• Risk-Aware and Non-Symmetric Interactions: MI deployment in Africainvolves risk-aware scenarios, acknowledging the uncertainties and chal-lenges unique to the continent. The interactions are non-symmetric, con-sidering the diverse contexts, economic landscapes, and technological in-frastructures across African nations.• Coopetition for Common Goals: While entities may compete in certainaspects of MI development, there is a recognition that collaboration onshared goals, such as addressing socio-economic challenges, improvinghealthcare, or promoting education, can lead to mutual benefits. Coope-tition arises as entities balance competition and cooperation to achieveoverarching objectives.• Partial Altruism and Partial Spite: Entities may exhibit partial altru-ism by contributing to initiatives that benefit the broader community orregion. Simultaneously, there might be elements of partial spite, wherecompetitive behaviors aim to gain advantages in specific domains. Thismix of cooperative and competitive motives characterizes the coopetitivenature of MI deployment.• Interactions Beyond Traditional Competition: The coopetitive Mean-Field-Type Game considers interactions that go beyond traditional competition.It involves collaborative efforts in areas like research and development,standard-setting, and addressing shared challenges, acknowledging thatcollective success can drive individual success.• Complex Interplay of Actors: The deployment of MI in Africa involvesa complex interplay of various actors, including governments, businesses,199academia, and communities. The interactions among these diverse stake-holders contribute to the coopetitive dynamics, shaping the trajectory ofMI development on the continent.Coopetitive MI in AfricaThere are over several hundred millions of people in Africa with diverse expe-riences in Agriculture, Breeding, Transformation, Trade, Traditional nutrition,Tradition, and Culture. Some of these Africans are champions in their fields.However, their innovations remain unknown and are often considered mysteri-ous due to a language barrier. They are not considered literate in the currentsystem, even though their audio-rich knowledge based on local experiences holdsgreat societal value. They are in fact audio-literate, not illiterate.Timadie’s Coopetition MI solution aims to place all Africans, particularlythose with unique knowledge, at the core of machine intelligence. Despite lan-guage barriers, our goal is to create direct audio-to-audio processing using high-quality, culture-aware datasets. Teams for each local language and bridgingteams connecting different languages form a coopetitive graph of machine in-telligence, not limited to text, audio, or video. Any knowledge-based learningmodel can join and enrich the Coopetition. We aim create value with theseaudio-rich knowledgeable people. The methods are : High-quality culture-sensitive & ethical audio dataset, Blockchainized audio token datasets, Bridgeto another local language: links. Examples include, but limited to, Tommo-SoDogon to Kenedougou Senoufo, Fon (from Benin) to Tommo-So Dogon, Audiodataset enhancement, Large Audio learning, Audio2Audio translation, Audioto visual (image/video) in a local context, Large vision learning, MultimodalLarge Learning, etc.10 Data in Africa: quality, information, modesTo develop an MI system tailored for the diverse needs of African people, acquir-ing and leveraging high-quality data becomes increasingly pivotal. The successand efficacy of MI in Africa are intricately tied to the richness and relevance ofthe datasets used for training and refining these systems. Gathering comprehen-sive and representative data from various regions, groups, ethnicities, languages,and socioeconomic backgrounds within the continent is essential to ensure thatthe resulting MI models are inclusive, unbiased, and culturally sensitive. Copy-pasting an MI from other cultures and other languages may not be efficient.Moreover, high-quality data is not solely about quantity but also about the ac-curacy, authenticity, and real-world applicability of the information collected. Itinvolves capturing the nuances of daily life, understanding regional contexts, andreflecting the dynamic nature of African societies. By prioritizing data quality,the development of local & context-aware MI in Africa can overcome challengesrelated to underrepresentation and biases, fostering a technology landscape thattruly resonates with the diverse realities of the continent. Additionally, a com-200mitment to ethical data practices is paramount. Safeguarding privacy, ensuringinformed consent, and promoting transparency in data collection processes areintegral aspects of building an MI infrastructure that aligns with African val-ues and respects individual rights. Ethical considerations not only enhance thereliability of MI applications but also foster trust among users, thereby encour-aging widespread adoption and acceptance of these technologies. In essence,recognizing the importance of high-quality data is foundational to the success-ful development and deployment of MI solutions in Africa. It is a collaborativeeffort that involves stakeholders from various sectors working together to curatedatasets that authentically represent the continent’s multifaceted landscape, ul-timately paving the way for MI systems that positively impact and empowerAfrican communities.10.1 Audio-rich languagesIn most African countries, the local population uses oral languages, producingaudios, speeches, voices, songs, and mimics, among others. These languagesare rich in audio content. Unfortunately, up to now, these audio resources havenot been fully utilized. A beneficial application of MI for the local populationin Africa could involve a direct conversion of audio from one oral language toanother. As an exercise we took the list of top 10 MI chatbots and asked simplequestions local African languages. We have tested with the three followingquestions:• How do you say ’thank you’ in Tommo-So Dogon?• Count from 1 to 5 in Tommo-So Dogon• How do you say ’what is your name?’ in Tommo-So Dogon ?The answers to these questions were not satisfactory as of December 2023. Thenwe tested the same 3 questions for 200 African languages.The answers were noteven close. This simple experience mirrors the situation faced by Africans whodon’t speak English. Many language models do not perform well for languageswith smaller numbers of speakers, especially African ones. The problem becomeseven more serious when we have that experience with those who cannot readand write but speak very well their mother tongue. Most of these languages areaudio-rich and currently there is MI adapted to local MI problems: translationin audio format from African language to another. Although there have beenefforts to include certain languages in MI models even when there is not muchdata available for training, these results show that the technology “really stillisn’t capturing our languages. This is due to the fact that the available resourceswhich are audio data are not exploited in these LLMs.10.2 Undefined terms to be clarified for machinesImagine a super-smart computer with data collected in Africa that can learnfrom pictures, words, and sounds all together. Sometimes, this computer can201make mistakes like believing things that aren’t true (delusion), creating thingsthat seem real but aren’t (illusion), making up stories that sound good but haveparts that aren’t real (confabulation), and even creating pictures or soundsthat are like dreams (hallucination). It might mix up where things came from(misattribution) or use words in new ways (semantic drift). It could make thingssound bigger or more important than they are (exaggeration) and tell storieswith fewer details than they need (semantic compression). Sometimes, it can’tfigure out what caused what (causal inference failure), and other times it makesthings that look almost real but not quite (perceptual uncanny valley). So,while it’s an amazing computer, it sometimes makes these mistakes that needto be defined properly, evaluated and corrected.The book in [3] addresses some of these questions.While these terms are unclear in MI, they are widely used in the literatureincluding in Africa. It does not help to clarify and demystify the methods usedin the building blocks of these MI technologies.10.3 Africa’s rich audio literacyYou want to engage in telemedicine in rural Africa, but face challenges due to lo-cal language barriers. Similarly, responding to a village’s call for adult businesseducation and facilitating communication between two farmers with 40 years oflocal agricultural experience, who cannot read or write, is hindered by languagebarriers. According to Statistica, it exceeds 30 percent of the 1.4 billion peoplein Africa. It is clear that an ethical, context-aware and culture-aware audio-to-audio MI technique could catalyze the inclusion of 400 millions of people inthese scenarios in Africa. In the vast landscape of technological evolution, Africaemerges as a reservoir of untapped potential. Defined by multifaceted richnessacross its 54 countries, the continent boasts a unique strength woven into thefabric of its diverse cultures - an inherent proficiency in audio literacy. Fromthe rhythmic beats of traditional music to the oral traditions passed throughgenerations, audio literacy is an integral part of African societies. This profi-ciency seamlessly extends into the digital domain, exemplified by the flourishingonline audio messaging platforms. We need to explore the transformative powerof leveraging Africa’s audio literacy as a catalyst for the advancement of MI inAfrica. With linguistic diversity and vibrant oral traditions, the continent holdsa reservoir of potential waiting to be tapped. The power of audio presents atransformative opportunity for technological innovation in various forms, includ-ing sounds, gestures, voices, images, mimics, signs, music, speeches, or videos.Several individuals in Africa, currently labeled as illiterate by the system, ex-tensively use mobile phones to send audio messages to their counterparts, andthese counterparts respond in audio in the same local language. This local lan-guage audio-to-audio communication has gained unprecedented momentum inseveral regions of Africa, generating a multitude of audio data. We realize thatthese individuals are audio communicators, they are in fact audio literate, cre-ating hope for inclusion through technology. Unfortunately, as of today, thereis no technology allowing audio-to-audio communication across different local202languages because these languages lack or minimally use conventional writingsystems and the current technologies are limited in textless audio processing . Itis imperative for us to transcend conventional modes of technology engagementand explore the possibilities that audio literacy unfurls. This article serves asa clarion call to stakeholders across the spectrum - policymakers, technologists,developers, educators, and innovators - to recognize and harness Africa’s richaudio literacy.Bridging the Tech Divide: Audio Literacy as the CatalystIn a continent rich with diverse cultures and languages, the intersection of MIand audio communication is not just a leap into the digital age but a catalystfor positive change. MI leveraging audio literacy is essential in Africa for itsability to bridge linguistic divides and promote inclusivity. With a multitudeof languages spoken across the continent, audio-based technology ensures thatinformation is accessible to a broader audience, transcending literacy barriers.This inclusivity becomes a driving force in education, healthcare, and commu-nication, fostering a more connected and empowered society. Additionally, asoral traditions play a significant role in many African cultures, embracing audioliteracy through MI respects and preserves these rich narratives while propellingcommunities into a digitally empowered future.Empowering Farmers Through Audio Knowledge SharingImagine a scenario where MI facilitates audio knowledge-sharing among farmersin local African languages like Tommo-So Dogon and Fon. In Tommo-So Dogon,farmers could exchange insights on sustainable agricultural practices throughvoice messages, discussing soil health or effective crop rotation. Simultaneously,in Fon, another group of farmers might share audio clips on traditional farmingtechniques or effective pest control methods. This audio-to-audio knowledgeexchange not only preserves cultural nuances but also creates a collaborativeplatform where diverse communities can benefit from each other’s expertise,fostering a stronger agricultural network across the continent.Developing Offline Audio-to-Audio Translation for FarmersMI can develop an offline, electricity-independent audio-to-audio translationtool for Tommo-So Dogon and Fon, catering to farmers’ specific needs. By in-corporating pre-trained models into portable, low-energy devices, such as solar-powered audio devices, farmers can access translations without the internet.These devices could utilize edge computing, processing information locally toovercome connectivity challenges. This textless translation model would betrained on a diverse dataset of both languages, understanding the unique pho-netics and nuances. This tailored solution empowers farmers with a user-friendlytool, ensuring vital knowledge exchange and collaboration, even in remote areaswith limited resources.203Unlocking the Potential of African Audio DatasetsA significant hurdle lies in the underutilization of the massive audio datasetsgenerated by Africans. Despite the wealth of linguistic diversity and culturalknowledge captured in these recordings, current MI implementations often over-look or lack access to these locally sourced datasets. Harnessing and incorpo-rating these valuable resources could substantially enhance the accuracy andeffectiveness of audio-to-audio translation tools, addressing the gap in inclusivetechnological solutions tailored to the specific needs of communities speakinglanguages like Tommo-So Dogon and Fon. The Imperative of Advanced AudioProcessing Tools Moreover, the need to develop advanced audio processing toolsbecomes evident in leveraging the rich diversity of African languages. Creatingaccurate audio-to-audio translation models demands sophisticated algorithmsthat can decipher subtle intonations, dialectical variations, and cultural con-text present in these recordings. Investing in the research and development ofcutting-edge audio processing tools is crucial to unlocking the full potential ofMI in catering to the unique linguistic landscape of Africa, facilitating effectiveknowledge exchange among farmers in languages like Tommo-So Dogon andFon.Closing the Technological Gap: Audio-Based MI for InclusivityIn Africa, it’s not the people who are late; it’s the technologies. The richness oforal traditions and audio literacy among diverse communities is palpable. How-ever, the technology landscape has yet to catch up. Remember, local languagesare rich in audio, but current technology struggles with audio processing. Ifthe technology transforms audio to text and translates and converts to audiothe translated text, it loses most of the interesting nuances since many of theselanguages lack a writing system. There’s a need for innovation to enable tech-nology to work directly with audio signals, mirroring how local African peoplecommunicate, without relying on text.Urgently, we need to bridge this gap by developing Audio-based Machine In-telligence. By harnessing the innate audio literacy of the people, we can createinclusive solutions that transcend linguistic barriers. It’s time for technologyto align with the vibrant mosaic of African cultures and empower communitiesthrough accessible, audio-driven advancements. Rethinking Tech Adoption: AnUrgent Call for Inclusive MI When we force people to adopt current MI technolo-gies, we create a greater tech divide in African society because the audio-literateare left out, and they constitute millions of people. However, if an audio-basedMI exists today, these millions of people could seize the opportunities to use itin their businesses, just as they started on audio messaging platforms withoutknowing how to read and write. All tech developers must consider this feedbackif the goal is to reduce the tech divide.It’s not because a piece of information is not available on the internet thatthis information doesn’t exist. It’s not because these millions of people don’twrite that their history doesn’t exist. They are audio-literate individuals that204technologies must learn from and integrate. In exploring MI in Africa, thecontinent’s abundant potential lies in its rich audio literacy. From traditionalrhythms to dynamic oral traditions, audio literacy is a transformative force.The call to leverage Africa’s audio proficiency for MI resonates as millions,labeled illiterate, actively engage in local language audio-to-audio communica-tion through smartphones. Despite challenges, the journey toward inclusiveMI involves recognizing cultural narratives, preserving oral traditions, and ad-dressing technological gaps. By embracing audio literacy, we pave the way fora future where technology aligns with African cultures, empowering communi-ties through accessible, audio-driven advancements, and reducing technologicaldivides.Machine Intelligence IntegrityMachine intelligence integrity can be understood as the intentional and system-atic approach to design, implement, and operate machine intelligence systemsin a manner that upholds safety, dignity, and ethical considerations throughouttheir lifecycle. It involves embedding ethical assessments into the core function-ing of the MI’s operating system, ensuring a continuous commitment to humanvalues and societal well-being.For lawyers, understanding machine intelligence integrity is crucial as it in-volves legal and ethical compliance, ensuring MI systems adhere to evolvingregulations and ethical standards. It also underscores the importance of legaloversight in addressing ethical challenges, safeguarding user rights, and con-tributing to the responsible and lawful deployment of MI technologies.Machine intelligence integrity is a foundational principle for developers, em-phasizing the conscientious and systematic approach to designing, developing,and maintaining MI systems with a commitment to ethical considerations andhuman values. It involves integrating ethical assessments directly into the coreof the MI operating system, ensuring that the system not only meets technicalstandards but also aligns with principles of safety, fairness, and dignity. Fordevelopers, machine integrity highlights the importance of proactive measuresduring the entire lifecycle of MI, promoting transparency, user-centric design,and continuous monitoring to address potential biases and ethical concerns. Byembracing machine integrity, developers contribute to building MI systems thatare not just technically robust but also ethically sound, fostering responsibleinnovation and positive societal impact.Machine intelligence integrity stands as a critical framework for governmentauthorities engaged in regulating and overseeing machine intelligence. Thisconcept underscores the intentional and systematic design, development, andongoing operation of MI systems with a focus on ethical considerations andalignment with human values. For government authorities, understanding andpromoting machine integrity means advocating for MI systems that prioritizesafety, fairness, and transparency, while also accommodating the evolving eth-ical landscape. It emphasizes the importance of regulations and oversight toensure that MI technologies are deployed responsibly, respecting legal frame-205works, and contributing positively to societal well-being. By incorporating ma-chine integrity into regulatory discussions, government authorities play a crucialrole in fostering an ethical and accountable MI ecosystem that serves the publicinterest.Machine Intelligence Integrity goes beyond the standardethicsScope and Focus:• Machine Intelligence Integrity: Primarily focuses on the intentional andsystematic design and operation of MI systems to ensure safety, dignity,and ethical considerations throughout their lifecycle.• Multi-Factor MI Ethics Over Time: Encompasses a broader scope, con-sidering various ethical factors that may evolve over time, including butnot limited to privacy, transparency, bias mitigation, and societal impact.Temporal Perspective:• Machine Intelligence Integrity: Emphasizes a continuous commitment toethical considerations throughout the entire lifespan of the MI system.• Multi-Factor MI Ethics Over Time: Acknowledges the dynamic nature ofethical concerns, recognizing that factors like societal norms and valuescan evolve, requiring ongoing ethical assessments and adaptations.Integration into Operating Systems:• Machine Intelligence Integrity: Involves embedding ethical assessmentsdirectly into the core functioning of the MI’s operating system, makingethical considerations an integral part of the MI’s behavior.• Multi-Factor MI Ethics Over Time: Requires periodic assessments andadjustments to ethical guidelines, often influencing MI design and decision-making processes but may not be as deeply integrated into the operatingsystem.Human Empowerment Emphasis:• Machine Intelligence Integrity: Posits that MI should serve the empower-ment of humanity, suggesting a focus on designing systems that contributepositively to human well-being.• Multi-Factor AI Ethics Over Time: Prioritizes ethical guidelines to ensureMI benefits humanity and avoids harm, but the emphasis may extendbeyond empowerment to encompass a broader set of considerations.206Proactive vs. Reactive Approach:• Machine Intelligence Integrity: Takes a proactive stance by designing MIsystems with integrity from the outset, aiming to prevent ethical issues.• Multi-Factor MI Ethics Over Time: May involve a more reactive approach,responding to emerging ethical concerns and adapting guidelines as soci-etal and technological landscapes evolve.Comprehensiveness of Ethical Factors:• Machine Integrity: Specifically addresses the intentional design and oper-ation of MI systems to ensure safety, dignity, and ethical considerations.• Multi-Factor MI Ethics Over Time: Encompasses a wider range of ethicalfactors, such as fairness, accountability, and explainability, recognizingthe multifaceted nature of ethical considerations in MI development anddeployment.How to Implement Machine Intelligence Integrity ?Implementing machine integrity involves a comprehensive approach to ensurethe ethical design, development, and operation of MI systems throughout theirlifecycle. This includes defining clear ethical principles aligned with humanvalues, embedding ethical assessments into the design process, using diverseand representative data to minimize biases, implementing transparency andexplainability mechanisms, continuous monitoring, user involvement, adaptivelearning, human-in-the-loop governance, privacy protection, regular audits, andresponsive post-deployment policies. The goal is to proactively address ethicalconcerns, promote fairness, and safeguard user privacy, fostering a system thatrespects human values and societal norms.Machine Intelligence integrity is a comprehensive framework for ensuringthe responsible design, development, and operation of MI systems. It involvesembedding multi-factor ethical considerations into the core functioning of MI op-erating systems, with a focus on upholding safety, fairness, and dignity in align-ment with human values. Multi-factor ethical adherence is assessed through aset of scalar values, each representing the adherence to specific ethical principles,while decision alignment is similarly evaluated with scalar values for alignmentto decision principles. The resulting multi-objective metric vector captures thenuanced evaluation of both ethical adherence and decision alignment, providinga holistic measure of machine integrity. This approach enables a dynamic andresponsive assessment that considers the evolving nature of ethics and humanvalues over time, making it a vital concept for fostering responsible MI practicesacross various domains.Part of the errors can be reduced using the following filter:207Knowns UnknownsBelieve asKnownKnownKnownsKnown Un-knownsBelief asUnknownUnknownKnownsUnknown Un-knownswhich is rewritten asKnowns UnknownsBelieve asKnownThings the MIknows it knowsThings the MIknows it doesn’tknowBelief asUnknownThings the MIdoesn’t know itknowsThings the MIdoesn’t know itdoesn’t knowIn dynamic game theory, a state-feedback strategy refers to a decision-making approach where the actions of a player are determined based on boththe current state of the system and the historical evolution of the game. Unlikeopen-loop strategies that depend solely on the initial conditions, state-feedbackstrategies enable players to adapt their actions dynamically in response to thechanging circumstances within the game. These strategies involve the use offeedback mechanisms, where information about the current state is continu-ously incorporated to adjust the player’s decisions, allowing for more flexibilityand responsiveness in complex, evolving dynamic environments.In Mean-Field-Type Game Theory (MFTG), a state-and-mean-field-typestrategy involves players making decisions based not only on their individualstate-action pairs but also on the distribution of all individual state-action pairs.In this strategy, each player’s action is influenced by both their own local infor-mation (state) and own individual mean-field (belief), and the collective behav-ior local information and local belief of the entire group. The mean-field typeterm here captures the distribution of own information and the distributionof others’ information. It allows to consider risk-aware decisions via variance,quantile, expectile, extremile, conditional value-at-risk, entropic value-at-risk,etc. State-and-mean-field-type feedback strategies are particularly relevant inrisk quantification in engineering.A state-and-mean-field-type feedback strategy can be defined as a measur-able function incorporating various factors to assess the integrity of machineintelligence. The measurable function is a result of:(a) Current time(b) Current high-quality data(c) Current extracted knowledge data(d) Current architecture design(e) Current training model(f) Current localized deployment model(g) Current privacy-preserving status(h)Current context-awareness(i) Current risk-awareness(j) Current localized culture-awareness208(k) States of the societies at the current time(l) Localized social norms of the societies at the corresponding states(m) Current beliefs about human valuesA state-and-mean-field-type feedback strategy for implementing machine in-telligence integrity involves creating a measurable function that considers factorssuch as initial and current quality data, time, knowledge, architecture design,training and deployment models, societal states, beliefs about human values,and cultural influences. This approach requires continuous monitoring, adap-tive learning, feedback loops, scenario planning, ethics committees, human-in-the-loop governance, privacy protection, regular audits, user involvement, andeducational initiatives. By integrating these elements, the strategy aims toproactively align the machine intelligence system with evolving ethical consider-ations, human values, and societal norms, fostering responsible and accountabledeployment throughout its lifecycle. The advantage of employing a state-and-mean-field-type feedback strategy in Machine Intelligence lies in its comprehen-sive and adaptive approach to managing complex, interconnected factors. Thisstrategy allows the system to continuously monitor and integrate various ele-ments such as initial and current data quality, temporal aspects, architecturaldesign, societal states, beliefs about human values, and cultural influences. Byincorporating a mean-field-type approach, which considers the influence of thecollective distribution of these factors, the system becomes more responsive toevolving ethical considerations and societal dynamics. This adaptability, cou-pled with proactive measures like scenario planning, regular audits, and userinvolvement, enhances the system’s ability to align with human values, addressbiases, and promote responsible innovation. Ultimately, the state-and-mean-field-type feedback strategy contributes to the development of machine intel-ligence systems that are not only technically robust but also ethically sound,fostering trust and positive societal impactAdaptive Machine IntegrityAdapting machine integrity to the evolution of ethics and human values involvesa proactive and dynamic approach. Continuous ethical assessment, feedbackmechanisms, and engagement with diverse stakeholders are essential for stayingaligned with changing societal perspectives. Keeping training data updated,forming ethics committees, and implementing scenario planning contribute toanticipating and addressing emerging ethical challenges. Utilizing version con-trol for iterative updates, ensuring legal and regulatory compliance, fosteringpublic discourse, and maintaining human oversight are crucial components inthe ongoing effort to navigate and integrate evolving ethical considerations intoMI systems. Education and awareness initiatives further support a collectivecommitment to upholding machine integrity in the face of shifting ethical land-scapes over time.20911 ConclusionThis survey article sheds light on the state of MI research and implementationin Africa. It underscores the diverse paths each country takes in embracingMI and other emerging technologies. The table highlighting countries withexplicit national MI strategies reflects a growing governmental awareness andcommitment in some regions. However, the article reveals a disparity betweenhigh-profile meetings held in luxurious venues and the actual impact on thelocal population, emphasizing the need for inclusive strategies. Moreover, theexamination of over 400 research articles highlights the prevalence of oversellingfindings and the importance of updating studies to practical MI algorithms.The ethical dimension emerges in the call to move beyond uniform rankings,considering factors like data transparency and concrete government actions fora more accurate assessment of the progress achieved by MI and its utilizationacross the continent. Ultimately, this survey article serves as a valuable resourcefor understanding the landscape of MI in Africa and advocating for meaningful,locally impactful advancements.Future workOur review of 400 articles in MI in Africa from 2018-2023 show the use of thesetechnologies in diverse settings. From Cabo Verde to the tip of Ethiopia, fromCape Town to Tunis via Madagascar, the informal sector appears to extensivelyharness machine intelligence.• In the future, local MI must assimilate this on-the-ground data to enhanceits knowledge, covering a significant portion of the informal sector.• The ability to learn audibly from one local language to another withoutrelying on a writing system opens up business opportunities for millionsof people who excel in their native languages despite being unable to reador write.• The multidimensional and multifactorial aspect of machine ethics, influ-enced by local culture and specific issues, should not be overlooked andmust be integrated from the design and algorithm training and learningphase.• Local MI capable of operating without internet or electricity could broadenaccess in rural areas, benefiting both the young and the elderly.• Developing strategic MI in underserved areas would help anticipate marketinteractions and the informal economy.• Quantifying the risks associated with strategic MI should be at the coreof design, taking into account ethical and cultural factors.210Examining closely 400 articles on MI in Africa, beyond the headlines, itemerges that• human learning, the learning of men and women, whether children oradults, takes a much more central place in discussions than machine learn-ing.• At the core is human learning, utilizing various tools, including machineassistance, as well as inspiration from nature.Early stage MI development topicsBelow we provide some selected topics and their early stage MI development inAfrican countries.MI Implementation for African Countries RealizationBlockchain and MI-enabled National ID Very Early StageBlockchain and MI-enabled Land Management Very Early StageAudio2Audio local languages Very Early StageText2Image local items Very Early StageImage2Audio local languages Very Early StageKnowledge-based MI Very Early StageRobust MI Infrastructure Very Early StageMI-enabled Cybersecurity Very Early StageAccessible MI Education Programs Very Early StageCollaborative MI Research Initiatives Very Early StageResponsible MI Policies and Regulations Very Early StageSector-wide Integration of MI Technologies Very Early StageSupportive Ecosystem for MI Startups Very Early StageTechnical MI Ethics Very Early StageNon-Technical MI Ethics Very Early StageEthical Considerations in MI Development Very Early StageData Accessibility, Privacy, and Security Very Early StageGovernment Investment in MI Initiatives Very Early StageGlobal Participation in MI Initiatives Very Early StageEthical Data Practices Very Early StageInclusive MI Integration Across Key Sectors Very Early StageInternational Partnerships for Knowledge Exchange Very Early StageFunding for MI Research and Development Very Early StageEmphasis on Responsible and Inclusive MI Practices Very Early StagePromotion of MI Education and Skilled Workforce Very Early StageDevelopment of MI Regulatory Frameworks Very Early StageStrategic MI Technology Integration in Key Sectors Very Early StageSupport for MI Startup Ecosystem Very Early StagePublic Awareness and Understanding of MI Very Early StageMI-driven Solutions for Socio-Economic Challenges Very Early StageInfrastructure for Advanced Data Analytics Very Early StageMI-driven Agricultural Innovations Very Early StageTelemedicine and Healthcare MI Applications Very Early StageMI in Renewable Energy Very Early StageMI in Environmental Conservation Very Early StageMI for Efficient Transportation and Logistics Very Early StageMI for Cities and Urban Planning Very Early StageMI in Education for Personalized Learning Very Early StageCommunity Engagement in MI Development Very Early StageMI for Disaster Response and Management Very Early StageMI-powered Financial Inclusion Initiatives Very Early StageTable 62: MI Implementation Topics for the Benefits of African CountriesConcrete actions for MI implementationsImplementing MI as a local public good for the population requires collabo-ration among traditional authorities, government bodies, educational institu-tions, industry stakeholders, and local communities. It cannot be reduced toa political announcement without any concrete followup. Regular assessments,transparency, and a commitment to inclusivity will contribute to the success-ful development of responsible and culture-aware MI for the benefit of African211societies and worldwide. Here are some concrete actions to be taken towardsimplementation:• Recommendation: Develop training programs for MI practitioners onethics and cultural sensitivity, catering to all Africans, including thosewho may be considered illiterate but are, in fact, highly aural literate.The equivalent of literacy for audio is often referred to as ”aural literacy”or ”audio literacy.” Aural literacy involves the ability to interpret, compre-hend, and create meaning from auditory information. This includes skillssuch as listening comprehension, recognizing and understanding variousaudio elements, and effectively using and producing audio content. Asour interactions with technology and media become more audio-centric,developing aural literacy becomes increasingly important for navigatingand understanding the information presented in audio formats.– Concrete Action: Integrate culture-aware ethical training modulesinto audio literacy MI courses and provide ongoing professional de-velopment opportunities for practitioners.• Recommendation: Invest in MI education programs at all educationallevels.– Concrete Action: Establish MI curriculum in both conventional andnon-conventional schools and universities, ensuring students are ex-posed to both technical and ethical aspects of MI adapted to localproblems.• Recommendation: Conduct public awareness campaigns on MI.– Concrete Action: Organize workshops, webinars, and communityevents to educate the public about MI, its applications, and potentialsocietal impacts.• Recommendation: Create audio MI literacy programs for policymakersand government officials.– Concrete Action: Offer specialized training for policymakers to un-derstand vision and audio MI technologies, implications, and policyconsiderations.• Recommendation: Establish MI ethics training for businesses and organi-zations.– Concrete Action: Provide workshops and resources to businesses toensure ethical MI practices in their operations.• Recommendation: Include MI in vocational and technical training pro-grams.212– Concrete Action: Integrate MI components into vocational trainingprograms to prepare a diverse workforce for emerging MI-related jobs.• Recommendation: Collaborate with international organizations for MIeducation initiatives.– Concrete Action: Form partnerships with global educational institu-tions and organizations to enhance MI education programs.• Recommendation: Develop a centralized platform for MI knowledge shar-ing.– Concrete Action: Create an online platform for sharing MI resources,research, and best practices among academia, industry, and policy-makers.• Recommendation: Translate MI educational materials into local languages.– Concrete Action: Ensure accessibility by translating MI educationalcontent into languages spoken in the region.• Recommendation: Foster partnerships between educational institutionsand industry.– Concrete Action: Facilitate internships, joint research projects, andindustry collaborations to bridge the gap between academia and in-dustry.• Recommendation: Establish MI mentorship programs.– Concrete Action: Pair experienced MI professionals with studentsand early-career professionals to provide guidance and support.• Recommendation: Include context-aware MI in continuous learning pro-grams for professionals.– Concrete Action: Develop short courses and workshops to keep pro-fessionals updated on the latest advancements and ethical consider-ations in MI.• Recommendation: Encourage research on the cultural impact of MI.– Concrete Action: Fund research projects that specifically investigatehow MI intersects with local cultures and societies.• Recommendation: Create a network of MI experts for knowledge ex-change.– Concrete Action: Facilitate regular conferences, forums, and net-working events for MI professionals to share experiences and insights.213• Recommendation: Establish a national context-aware MI day to celebrateachievements and raise awareness.– Concrete Action: Designate a day to showcase MI projects, hostevents, and engage the rural & urban area public in discussions aboutMI’s role in society.• Recommendation: Incorporate context & cultural considerations into MIdevelopment guidelines.– Concrete Action: Work with traditionalists, ethicists, cultural ex-perts, and technologists to create MI development guidelines thatreflect cultural values.• Recommendation: Establish MI ethics review boards.– Concrete Action: Form independent boards to evaluate MI projectsfor cultural & ethical considerations, ensuring diverse representationand transparency.• Recommendation: Develop guidelines for responsible MI research.– Concrete Action: Create a set of guidelines for researchers to follow,addressing context-aware ethical considerations and potential societalimpacts.• Recommendation: Conduct regular ethical audits of MI systems.– Concrete Action: Implement periodic reviews of MI systems to iden-tify and address ethical concerns, with findings made public.• Recommendation: Encourage diverse teams in MI development.– Concrete Action: Promote diversity and inclusion in MI teams, en-suring representation from various backgrounds and perspectives.• Recommendation: Integrate bias detection tools into MI development pro-cesses.– Concrete Action: Require developers to use tools that identify andmitigate biases in MI algorithms during the development phase.• Recommendation: Establish clear guidelines for MI transparency.– Concrete Action: Define standards for transparency in local MI sys-tems, ensuring clear communication about how decisions are made.• Recommendation: Implement continuous training on ethical MI for devel-opers.– Concrete Action: Mandate ongoing training for MI developers to stayupdated on ethical considerations and best practices.214• Recommendation: Encourage the adoption of MI ethics certifications.– Concrete Action: Support the development and recognition of certi-fications that highlight adherence to ethical MI practices.• Recommendation: Establish mechanisms for reporting MI ethical con-cerns.– Concrete Action: Create channels for reporting ethical concerns re-lated to MI applications, ensuring a responsive and accountable sys-tem.• Recommendation: Foster a culture of responsible MI innovation.– Concrete Action: Recognize and reward organizations and individ-uals who prioritize responsible MI development in competitions andawards.• Recommendation: Include ethical considerations in MI project fundingcriteria.– Concrete Action: Funding agencies should prioritize projects thatadhere to ethical guidelines and promote responsible MI practices.• Recommendation: Collaborate with international organizations on ethicalMI standards.– Concrete Action: Participate in global discussions to contribute tothe development of ethical MI standards and adopt them nationally.• Recommendation: Regularly update ethical guidelines to reflect evolvingtechnologies.– Concrete Action: Establish a process for periodic review and updatesof ethical guidelines to keep pace with technological advancements.• Recommendation: Establish an independent body for monitoring MI ethics.– Concrete Action: Form a national or regional body responsible formonitoring and enforcing ethical and culture-aware MI practices,with the power to investigate and penalize non-compliance.• Recommendation: Incorporate community input into MI decision-making.– Concrete Action: Implement mechanisms for seeking public inputand feedback on MI projects, ensuring community perspectives areconsidered.• Recommendation: Conduct impact assessments for MI implementations.215– Concrete Action: Require organizations to conduct thorough impactassessments before deploying MI systems, assessing potential social,economic, and cultural implications.• Recommendation: Support community-based MI initiatives.– Concrete Action: Provide grants and resources to community orga-nizations working on MI projects that address local challenges andpriorities.• Recommendation: Establish community-led MI forums.– Concrete Action: Facilitate regular community forums where resi-dents can voice concerns, ask questions, and engage in discussionsabout MI applications.• Recommendation: Create MI literacy programs for community leaders.– Concrete Action: Develop tailored programs to educate communityleaders about MI, enabling them to advocate for responsible MI de-velopment.• Recommendation: Collaborate with traditional leaders and elders on MIinitiatives.– Concrete Action: Engage with local traditional leaders to incorporatecultural wisdom and perspectives into MI projects.• Recommendation: Develop accessible MI interfaces for diverse popula-tions.– Concrete Action: Ensure MI interfaces are user-friendly and accessi-ble to diverse populations, considering language, literacy levels, anddigital access.• Recommendation: Promote gender-inclusive MI development.– Concrete Action: Encourage the participation of women in MI projectsand ensure gender considerations are integrated into the developmentprocess.• Recommendation: Establish partnerships with civil society organizations.– Concrete Action: Collaborate with local teams, associations and civilsociety groups to ensure MI projects align with societal values andaddress community needs.• Recommendation: Conduct MI impact assessments on marginalized com-munities.216– Concrete Action: Prioritize assessments on how MI implementationsmay affect marginalized or vulnerable communities, addressing po-tential risks.• Recommendation: Encourage local innovation in MI for societal chal-lenges.– Concrete Action: Create funding opportunities and innovation chal-lenges specifically aimed at addressing local societal challenges throughMI.• Recommendation: Establish mechanisms for MI project feedback and im-provement.– Concrete Action: Create channels for ongoing feedback from com-munities impacted by MI projects, enabling continuous improvementbased on real-world experiences.• Recommendation: Foster cultural preservation through MI initiatives.– Concrete Action: Support projects that leverage MI to preserve andpromote cultural heritage, languages, and traditions.• Recommendation: Encourage MI for social inclusion and accessibility.– Concrete Action: Prioritize MI applications that enhance social in-clusion and accessibility for people with disabilities, elderly popula-tions, and other marginalized groups.• Recommendation: Establish a national or regional MI advisory councilthat works with an MI global alliance.– Concrete Action: Form a council comprising representatives fromdiverse sectors, including informal sector, academia, industry, civilsociety, and government, to provide guidance on MI developmentand its societal impact.The sections within this survey have unfolded a narrative that goes be-yond the mere exploration of technological landscapes; it is a testament to thedynamic interplay between MI and the rich, diverse cultures of the African con-tinent. In our journey through North Africa, Southern Africa, Middle Africa,West Africa, and East Africa, we’ve witnessed the transformative power of MI,transcending geographical boundaries to leave an indelible mark on the informaleconomy, agriculture, art and music, and societies. Yet, embedded within theseadvancements is a constant reminder of the importance of cultural sensitivity,ethical considerations, and the need to harness the unique strengths of Africa’saudio literacy. The call to action resonates even more strongly now - the im-perative to leverage Africa’s audio literacy as a cornerstone for MI innovation.In the era of digital communication, where the auditory sense is increasinglybecoming a dominant mode of interaction, the significance of audio literacy217cannot be overstated. This article extends a resounding call to the global techcommunity, urging a collective acknowledgment and appreciation for the powerof audio in shaping our technological landscape. The African experience, withits innate proficiency in audio literacy, serves as an inspiring model for theworld. The fusion of MI and audio literacy has the potential to break downbarriers, foster inclusivity, and create technologies that resonate with people ona deeper, more human level. To the tech innovators, developers, and engineersworldwide, consider this an invitation to reevaluate the role of audio in your cre-ations. Imagine a future where technology not only speaks to users but engageswith them in a language that transcends written words - a language embeddedin the rhythms, intonations, and nuances of audio. As we champion the cause ofaudio literacy, let us embark on a journey to make technology more accessible,inclusive, and culturally sensitive. Whether you’re designing virtual assistants,educational tools, or entertainment platforms, integrating audio literacy into thefabric of your innovations can pave the way for a more universally understoodand appreciated technological landscape. This is a call to infuse cultural rich-ness into the very code and algorithms that shape our digital experiences. Letaudio become a bridge that connects diverse communities, ensuring that tech-nology resonates with people from various linguistic and cultural backgrounds.In the spirit of coopetition, where collaboration coexists with competition, letus share insights, methodologies, and best practices for incorporating audio lit-eracy into tech solutions. By doing so, we can collectively contribute to a globaltechnological ecosystem that celebrates the richness of human expression in allits auditory dimensions. This article is also an invitation to envision a futurewhere the dialogue between humans and machines extends beyond the visualand written to embrace the profound power of sound. The path ahead is one ofcollaboration, innovation, and a shared commitment to shaping a technologicalfuture that is not just intelligent but also deeply attuned to the audio path ofhumanity.AcknowledgmentThe work of Hamidou Tembine is supported by Timadie grant on Mean-Field-Type Game Theory for Machine Intelligence in Underserved Areas.References[1] Jean Epstein, l’intelligence d’une machine, les classiques, Publisher: LeEditions Jacques Melot, 1946.[2] Epstein, J. (1946/2014). The intelligence of a machine (C. Wall-Romana,Trans.). Univocal Publishing.[3] Modibo Bouaré, Sidy Danioko, Mariam Dembélé, Abdoulaye Baniré Di-allo, Boubacar Diallo, Abdoulaye Diarra, Bourama Doumbia, Ndeye Sis-218soko Molinier, Astou Nathalie Sidibé, Allahsera Auguste Tapo, HamidouTembine : L’intelligence machine en Afrique en 20 questions, 181 pages,Sawa Editions, National Library of Mali, June 2023, ISBN : 978 9995294069[4] Allahsera Auguste Tapo, Ali Traoré, Sidy Danioko, Hamidou Tembine :Machine Intelligence in Africa: a survey, DSAI Conference 2024.[5] Manzoor Khan and Hamidou Tembine, Meta-Learning for Realizing Self-xManagement of Future Networks, IEEE Access Journal, vol.5, pp. 19072- 19083, August , 2017[6] H. Tembine, Deep Learning Meets Game Theory: Bregman-based Algo-rithms for Interactive Deep Generative Adversarial Networks, IEEE Trans-actions on Cybernetics, 14 pages, volume 50 , Issue: 3 , March, pp. 1132 -1145, 2020[7] J. Gao and H. Tembine, Distributionally Robust Games for Deep Gen-erative Learning, IEEE World Congress on Computational IntelligenceWindsor Convention Centre, Rio de Janeiro, Brazil 08-13 July, 2018[8] J. Gao and H. Tembine, Distributionally Robust Games: WassersteinMetric, International Joint Conference on Neural Networks (IJCNN), Riode Janeiro, Brazil, July, 2018[9] M. Khan and H. Tembine and A. Vasilakos, Game Dynamics and Cost ofLearning in Heterogeneous 4G Networks, IEEE Journal on Selected Areasin Communications, vol 30, 1, pp 198 - 213, January , 2012[10] Arakpogun, E., Elsahn, Z., Olan, F., Elsahn, F.: Artificial intelligence inAfrica: challenges and opportunities. pp. 375-388 (2020) In: The FourthIndustrial Revolution: Implementation of Artificial Intelligence for Grow-ing Business Success. Studies in Computational Intelligence . Springer,Cham, Switzerland, pp. 375-388. ISBN 9783030627980; 9783030627959;9783030627966[11] Gwagwa, A., Kraemer-Mbula, E., Rizk, N., Rutenberg, I., De Beer, J.:Artificial intelligence (AI) deployments in Africa: benefits, challenges andpolicy dimensions. African J. Inf. Commun. 26, 1 - 28 (2020).[12] Thomas Hervé Mboa Nkoudou, We need a decolonized appropriationof AI in Africa, Nature Human Behaviour, Pub Date: 2023-11-20,DOI:10.1038/s41562-023-01741-3[13] Suzuki, S. We need a culturally aware approach to AI. Nat Hum Behav ,1816 - 1817 (2023). https://doi.org/10.1038/s41562-023-01738-y[14] Absalom E Ezugwu, Olaide N Oyelade, Abiodun M Ikotun, Jeffery OAgushaka, Yuh-Shan Ho, Machine Learning Research Trends in Africa:A 30 Years Overview with Bibliometric Analysis Review, Archives of Com-putational Methods in Engineering, 1-31, 2023219[15] Chimango Nyasulu, Awa Diattara, Assitan Traore, Cheikh Ba, Enhancingfarmers productivity through IoT and machine learning: a state-of-the-artreview of recent trends in Africa , Research in Computer Science and ItsApplications: 11th International Conference, CNRIA 2021, Virtual Event,June 17-19, 2021, Proceedings, 113-124, 2021[16] Alupo, Charity Delmus, Daniel Omeiza, and David Vernon. ”Realizing thePotential of AI in Africa: It All Turns on Trust.” In Towards TrustworthyArtificial Intelligent Systems, pp. 179-192. Cham: Springer InternationalPublishing, 2022.[17] Ndonye, Michael M. ”Mass-mediated feminist scholarship failure in Africa:normalised body-objectification as artificial intelligence (AI).” Editon Con-sortium Journal of Media and Communication Studies 1, no. 1 (2019): 1-8.[18] Mathenge, Wanjiku Ciku. ”Artificial intelligence for diabetic retinopathyscreening in Africa.” The Lancet Digital Health 1, no. 1 (2019): e6-e7.[19] Sallstrom, Laura, Olive Morris, and Halak Mehta. ”Artificial intelligence inAfrica’s healthcare: Ethical considerations.” ORF Issue Brief 312 (2019).[20] Rutenberg, Isaac, Arthur Gwagwa, and Melissa Omino. ”Use and Impact ofArtificial Intelligence on Climate Change Adaptation in Africa.” In AfricanHandbook of Climate Change Adaptation, pp. 1107-1126. Cham: SpringerInternational Publishing, 2021.[21] Sey, Araba, Gabriella Razzano, Andrew Rens, and Shamira Ahmed. ”Map-ping policy and capacity for artificial intelligence for development inAfrica.” (2021).[22] Couchoro, Mawuli K., Koffi Sodokin, and Moubarak Koriko. ”Informa-tion and communication technologies, artificial intelligence, and the fightagainst money laundering in Africa.” Strategic Change 30, no. 3 (2021):281-291.[23] Pantserev, Konstantin A. ”Malicious use of artificial intelligence in Sub-Saharan Africa: Challenges for Pan-African cybersecurity.” Vestnik RUDN.International Relations 22, no. 2 (2022): 288-302.[24] Neupane, Bhanu, and Prateek Sibal. Artificial intelligence needs assessmentsurvey in Africa. UNESCO Publishing, 2021.[25] Botwe, B.O., Akudjedu, T.N., Antwi, W.K., Rockson, P., Mkoloma, S.S.,Balogun, E.O., Elshami, W., Bwambale, J., Barare, C., Mdletshe, S. andYao, B., 2021. The integration of artificial intelligence in medical imag-ing practice: perspectives of African radiographers. Radiography, 27(3),pp.861-866.220[26] Kothari, A. and Cruikshank, S.A., 2022. Artificial intelligence and jour-nalism: An Agenda for journalism research in Africa. African JournalismStudies, 43(1), pp.17-33.[27] Kiemde, Sountongnoma Martial Anicet, and Ahmed Dooguy Kora. ”To-wards an ethics of AI in Africa: rule of education.” AI and Ethics (2022):1-6.[28] Nayebare, Micheal. ”Artificial intelligence policies in Africa over the nextfive years.” XRDS: Crossroads, The ACM Magazine for Students 26, no. 2(2019): 50-54.[29] Kohnert, Dirk. ”Machine ethics and African identities: Perspectives of ar-tificial intelligence in Africa.” Available at SSRN 4163096 (2022).[30] Otaigbe, Idemudia. ”Scaling up artificial intelligence to curb infectious dis-eases in Africa.” Frontiers in Digital Health 4 (2022): 1030427.[31] Ade-Ibijola, A. and Okonkwo, C., 2023. Artificial Intelligence in Africa:Emerging Challenges. In Responsible AI in Africa: Challenges and Oppor-tunities (pp. 101-117). Cham: Springer International Publishing.[32] Gravett, Willem H. ”Digital coloniser? China and artificial intelligence inAfrica.” In Survival December 2020 - January 2021: A World After Trump,pp. 153-177. Routledge, 2023[33] Hermann Ndoya, Donald Okere, Marie laure Belomo, Melissa Atangana,Does ICTs decrease the spread of the informal economy in Africa? Telecom-munications Policy, Volume 47, Issue 2, 2023, 102485, ISSN 0308-5961,https://doi.org/10.1016/j.telpol.2022.102485.[34] Plantinga, Paul and Shilongo, Kristophina and Mudongo, Oarabile andUmubyeyi, Angelique and Gastrow, Michael and Razzano, Gabriella, Re-sponsible artificial intelligence in Africa: Towards policy learning, 2023,SocArXiv[35] Jaldi, Abdessalam, Artificial Intelligence Revolution in Africa: EconomicOpportunities and Legal Challenges, 2023[36] Gwagwa, Arthur and Kazim, Emre and Kachidza, Patti and Hilliard, Air-lie and Siminyu, Kathleen and Smith, Matthew and Shawe-Taylor, John,Road map for research on responsible artificial intelligence for development(AI4D) in African countries: The case study of agriculture, Patterns, 2, 12,2021, Elsevier[37] Adeshina, Steve A and Aina, Oluwatomisin, The Role of AI in SDG: AnAfrican Perspective, The Ethics of Artificial Intelligence for the SustainableDevelopment Goals, 133–143, 2023, Springer.221[38] Delphin Kamanda Espoir, Regret Sunge, Tii Nchofoung, Andrew AdewaleAlola, Analysing the drivers of ecological footprint in Africa with machinelearning algorithm, Environmental Impact Assessment Review 104, 107332,2024[39] Enock Mintah Ampaw, Junwu Chai, Yuguo Jiang, Adjei Peter Darko,Kwame Simpe Ofori, Rethinking small-scale gold mining in Ghana: A holygrail for environmental stewardship and sustainability. Journal of CleanerProduction, 140683, 2024[40] Wiston Forero-Corba, Francisca Negre Bennasar, Techniques and applica-tions of Machine Learning and Artificial Intelligence in education: a sys-tematic review, Revista Iberoamericana de Educacion a Distancia 27 (1),209-238, 2024[41] Clinton Aigbavboa, Douglas Aghimien, Emmanuel Aghimien, John Aliu,Ayodeji Oke, Daniel WM Chan, Machine Learning-Based Sensitivity Anal-ysis of the Strategies for Construction Digitalisation, Preprint 2024[42] Rebecca Peters, Jurgen Berlekamp, Charles Kabiri, Beth A Kaplin, Kle-ment Tockner, Christiane Zarfl, Sustainable pathways towards universalrenewable electricity access in Africa, Nature Reviews Earth & Environ-ment, 1-15, 2024[43] Prince Amoah, Gabriel Eweje, Examining the social sustainability strate-gies of multinational mining companies in a developing country, Social Re-sponsibility Journal 20 (3), 568-584, 2024[44] Siham Eddamiri, Fatima Zahra Bassine, Victor Ongoma, Terence EpuleEpule, Abdelghani Chehbouni, An automatic ensemble machine learningfor wheat yield prediction in Africa, Multimedia Tools and Applications,1-27, 2024[45] Jesse Gilbert, Jeffrey NA Aryee, Mary Jessie Adjei, Machine-LearningApproaches for Assessing Aerosol Optical Depth (AOD) in Ghana, WestAfrica, EarthArXiv, 2024.[46] O Olawale Awe, Kim Love, Eric A Vance, Teaching Data Science in AfricaVia Online Team-Based Learning, Sustainable Statistical and Data ScienceMethods and Practices: Reports from LISA 2020 Global Network, Ghana,2022, 409-415, 2024[47] Sokhna Fatou Seck Sarr, Artificial Intelligence and Fact Checking in Africa:Between Logic of Dependency and the Limits of Automation, ESI Preprints25, 374-374, 2024[48] Allen Munoriyarwa, Sarah Chiumbu, Artificial intelligence skepticism innews production: The case of South Africa’s mainstream news organiza-tions , Global Journalism in Comparative Perspective, 117-131, 2024222[49] Anna Bon, Francis Saa-Dittoh, Hans Akkermans, Bridging the digital di-vide, Book chapter in Introduction to Digital Humanism, Editors: HannesWerthner, Carlo Ghezzi Jeff Kramer, Julian Nida-Rumelin, Bashar Nu-seibeh, Erich Prem, 283, Springer 2024[50] Nick Bryan-Kinns, Bingyuan Zhang, Songyan Zhao, Berker Banar, Explor-ing Variational Auto-encoder Architectures, Configurations, and Datasetsfor Generative Music Explainable AI, Machine Intelligence Research 21 (1),29-45, 2024[51] Kevin Mallinger, Ricardo Baeza-Yates. Responsible AI in Farming: AMulti-Criteria Framework for Sustainable Technology Design , Applied Sci-ences 14 (1), 437, 2024[52] Ussen Kimanuka, Ciira wa Maina, Osman Buyuk, Speech recognitiondatasets for low-resource Congolese languages, Data in Brief 52, 109796,2024[53] Johannes Brozovsky, Nathalie Labonnote, Olli Vigren, Digital technologiesin architecture, engineering, and construction, Automation in Construction158, 105212, 2024[54] Adam M Chekroud, Matt Hawrilenko, Hieronimus Loho, Julia Bondar,Ralitza Gueorguieva, Alkomiet Hasan, Joseph Kambeitz, Philip R Cor-lett, Nikolaos Koutsouleris, Harlan M Krumholz, John H Krystal, MartinPaulus, Illusory generalizability of clinical prediction models, Science 383(6679), 164-167, 2024[55] Abdullah Al Maruf, Fahima Khanam, Md Mahmudul Haque, Zakaria Ma-sud Jiyad, Firoj Mridha, Zeyar Aung, Challenges and Opportunities ofText-based Emotion Detection: A Survey, IEEE Access, 2024[56] Jian Gao, Deep Chakraborty, Hamidou Tembine, Olaitan Olaleye: Non-parallel Emotional Speech Conversion, Proc. Interspeech 2019, 2858-2862.[57] Abdul-Hakeem Omotayo, Mai Gamal, Eman Ehab, Gbetondji Dovonon,Zainab Akinjobi, Ismaila Lukman, Houcemeddine Turki, Mahmod Ab-dien, Idriss Tondji, Abigail Oppong, Yvan Pimi, Karim Gamal, Ro’ya-CV4Africa, Mennatullah Siam, Towards a Better Understanding of theComputer Vision Research Community in Africa, EAAMO ’23: Proceed-ings of the 3rd ACM Conference on Equity and Access in Algorithms,Mechanisms, and Optimization, October 2023 No.: 1, Pages 1-12[58] Abdul-Hakeem Omotayo, Ashery Mbilinyi, Lukman Ismaila, Houcemed-dine Turki, Mahmoud Abdien, Karim Gamal, Idriss Tondji, Yvan Pimi,Naome A. Etori, Marwa M. Matar, Clifford Broni-Bediako, Abigail Op-pong, Mai Gamal, Eman Ehab, Gbetondji Dovonon, Zainab Akinjobi,Daniel Ajisafe, Oluwabukola G. Adegboro, Mennatullah Siam, A Surveyon African Computer Vision Datasets,Topics and Researchers, Preprint2024223[59] Mazari, F. Z., Bouchaour, R. (2023). Economic intelligence: a mechanismto support and develop Small and Medium Enterprises in Algeria. VolumeIX, number 02.[60] Algerian Press Service. Presentation of the National Artificial IntelligenceStrategy 2020-2030.[61] Algerian Press Service. (n.d.). Higher education: First House of ArtificialIntelligence opens at Algiers University.[62] ENSIA. Retrieved from https://www.ensia.edu.dz/[63] Bennihi, Aymen Salah, Bouriche, Lahcene, Schneider, Friedrich. (2021).The informal economy in Algeria: New insights using the MIMIC approachand the interaction with the formal economy. Economic Analysis and Pol-icy, 72, 470-491.[64] Government of Egypt. National MI strategy of Egypt. Retrieved fromhttps://ai.gov.eg/[65] Egyptian Journal of Artificial Intelligence.. Retrieved from https://ejai.journals.ekb.eg/[66] Ragheb, Mohamed A., Tantawi, Passent, Farouk, Nevien, Hatata, Ahmed.(2022). International Journal of Higher Education Management, 8 (2).[67] Gaafar, Ali Sayed Mohamed, Allah, Hebat. (2020). Artificial intelligence inEgyptian tourism companies: Implementation and perception. Journal ofAssociation of Arab Universities for Tourism and Hospitality, 18 (1), 66-78.[68] Abu Aeshah, Ahmed Naji Ahmed. (2022). Artificial intelligence applicationin the adoption of e-commerce in small and medium enterprises in Libya.Journal of University Studies for Inclusive Research, 6 (11), 8309-3826.[69] Masoud, Najeb. (2014). Predicting direction of stock prices index move-ment using artificial neural networks: The case of Libyan financial market.British Journal of Economics, Management & Trade, 4 (4), 597-619.[70] Alarbi, Ali AA, Strickland, Dani, Blanchard, Richard. (2019). MI conceptsfor Demand Side Shedding Management in Libya. In 2019 8th InternationalConference on Renewable Energy Research and Applications (ICRERA)(pp. 505-510).[71] Alsharif, Ibrahem, Hebrisha, Hamza Emhemed, Ahmed, Abdussalam Ali.(2023). A Comprehensive Review Towards Libyan Smart Cities. AfricanJournal of Advanced Pure and Applied Sciences (AJAPAS), 15-22.[72] Oubibi, Mohamed, Zhou, Yueliang, Oubibi, Ayoub, Fute, Antony, Saleem,Atif. (2022). The challenges and opportunities for developing the use ofdata and artificial intelligence (AI) in North Africa: Case of Morocco. InInternational Conference on Digital Technologies and Applications (pp. 80-90).224[73] Bensalah, Mounir. (2021). Toward an ethical code of artificial intelligenceand human rights in Morocco. Arribat-International Journal of HumanRights, 1(2), 187-203.[74] Naim, Ayman, Aaroud, Abdessadek, Akodadi, Khalid, El Hachimi,Chouaib. (2021). A fully AI-based system to automate water meter datacollection in Morocco country. Array, 10, 100056.[75] Ouassou, El houssin, Taya, Hafsa. (2022). Forecasting regional tourismdemand in Morocco from traditional and AI-based methods to ensemblemodeling. Forecasting, 4 (2), 420-437.[76] Hajji, Said. (2023). Educational Sovereignty and Artificial IntelligenceChallenges: The Case of Morocco. In Artificial Intelligence in Higher Ed-ucation and Scientific Research: Future Development (pp. 101-116).[77] Mkik, Marouane, Mkik, Salwa. (2023). Acceptability Aspects of ArtificialIntelligence in Morocco: Managerial and Theoretical Contributions. In In-ternational Conference on Digital Technologies and Applications (pp. 65-74).[78] Hniche, Omar, Saadane, Rachid. (2023). Challenges and Opportunities ofIntegrating MI with IFRS in Accounting Systems: Insights from Morocco.In Proceedings of the 6th International Conference on Networking, Intelli-gent Systems & Security (pp. 1-9).[79] Mohammed, Musaab AA, Khleel, Nasraldeen AA, Szabó, Norbert P, Szűcs,Péter. (2023). Modeling of groundwater quality index by using artificialintelligence algorithms in northern Khartoum State, Sudan.Modeling EarthSystems and Environment, 9 (2), 2501-2516.[80] Hussein, Hussein Ali, Mahmoud, Mukhtar ME, Eisa, Haroun A. (2023).Performance Evaluation of ARIMA and FB-Prophet Forecasting Methodsin the Context of Endemic Diseases: A Case Study of Gedaref State inSudan. EMI Endorsed Transactions on Smart Cities, 7 (2), e1-e1.[81] National Agency for the Promotion of Scientific Re-search. (2018). National Artificial Intelligence Strat-egy: Unlocking Tunisia’s capacity potential. www.anpr.tn/national-ai-strategy-unlocking-tunisias-capabilities-potential[82] Bennani, F. (2018). Tunisia - Innovative Startups and SMEs Project(P167380): Concept Project Information Document. The World Bank.[83] Chatbri, Houssem, Jemmali, Sahbi, Hannachi, Salah. (2019). Empoweringthe Tunisian textile industry with artificial intelligence. Preprint.[84] Frikha, Tarek, Ktari, Jalel, Ben Amor, Nader, Chaabane, Faten, Hamdi,Monia, Denguir, Fehmi, Hamam, Habib. (2023). Low Power Blockchain inIndustry 4.0 Case Study: Water Management in Tunisia. Journal of SignalProcessing Systems, 1-15.225[85] Jammeli, Haifa, Ksantini, Riadh, Ben Abdelaziz, Fouad, Masri, Hatem.(2021). Sequential artificial intelligence models to forecast urban solid wastein the city of Sousse, Tunisia. IEEE Transactions on Engineering Manage-ment.[86] Messaoudi, Abir, Haddad, Hatem, Fourati, Chayma, BenHaj Hmida, Ay-men Ben Elhaj, Graiet, Mohamed. (2021). Tunisian dialectal end-to-endspeech recognition based on deepspeech. Procedia Computer Science, 189,183-190.[87] Smartson P Nyoni, Thabani Nyoni, Forecasting Infant Mortality Rate inCameroon Using Artificial Neural Networks, International Research Jour-nal of Innovations in Engineering and Technology 5 (3), 612, 2021[88] Zhi Zhen Qin, Melissa S Sander, Bishwa Rai, Collins N Titahong, SantatSudrungrot, Sylvain N Laah, Lal Mani Adhikari, E Jane Carter, LekhaPuri, Andrew J Codlin, Jacob Creswell, Using artificial intelligence to readchest radiographs for tuberculosis detection: A multi-site evaluation of thediagnostic accuracy of three deep learning systems ,Scientific reports 9 (1),15000, 2019[89] AS Andigema, NNT Cyrielle, KP Landry, AJ Vladimir, AI in the Manage-ment of HIV: Case Study Cameroon , Int J Virol AIDS 10, 089, 2023.[90] Inès Baleydier, Pierre Vassilakos, Roser Vinals, Ania Wisniak, Bruno Ken-fack, Jovanny Tsuala Fouogue, George Enownchong Enow Orock, SophieLemoupa Makajio, Evelyn Foguem Tincho, Manuela Undurraga, MagaliCattin, Solomzi Makohliso, Klaus SchÖnenberger, Alain Gervaix, Jean-Philippe Thiran, Patrick Petignat, Study protocol for a two-site clinicaltrial to validate a smartphone-based artificial intelligence classifier identi-fying cervical precancer and cancer in HPV-positive women in Cameroon ,Plos one 16 (12), e0260776, 2021.[91] Edgard Ndassimba, Nadege Gladys Ndassimba, Ghislain Mervyl Kossin-gou, Samuel Ouya, Smart agriculture solution based on IoT and TVWSfor arid regions of the Central African Republic, Networking, IntelligentSystems and Security: Proceedings of NISS 2021, 653-666, 2022[92] Galina Alova, Philipp A Trotter, Alex Money, A machine-learning approachto predicting Africa’s electricity mix based on planned power plants andtheir chances of success , Nature Energy 6 (2), 158-166, 2021.[93] Thomas Pellegrini, Deep-learning-based central African primate speciesclassification with MixUp and SpecAugment, Interspeech, 2021.[94] Ghislain Mervyl Saint-Juste Kossingou, Bessan Melckior Dégboé, SamuelOuya, Gervais Mendy, Proposal of algorithms to make up for lost schooland university years in post-conflict African countries in the face of Covid-19: case of the Central African Republic , International Journal of MobileLearning and Organisation 16 (4), 507-525, 2022.226[95] George Wu, Bin Li, Application of Machine Learning on the Diag-nosis of 18 Common Pediatric Diseases in Central African Repub-lic, In: Saroj Hiranwal and Garima Mathur (eds), Artificial Intelli-gence and Communication Technologies, SCRS, India, 2022, pp. 521-525.https://doi.org/10.52458/978-81-955020-5-9-50[96] Smartson P Nyoni, Tatenda A Chihoho, Thabani Nyoni, Forecasting TotalFertility Rate in Chad Using a Machine Learning Approach, InternationalResearch Journal of Innovations in Engineering and Technology 5 (8), 155,2021.[97] Kim-Ndor Djimadoumngar, Using machine learning algorithms to analyzeremote sensing and ground-truth Lake Chad’s level data, EGUsphere, 1-22,2022.[98] Jannis Hoch, Sophie Pieternel de Bruin, Halvard Buhaug, Nina vonUexkull, Rens van Beek, Niko Wanders, Projecting armed conflict risk inAfrica towards 2050 along the Shared Socio-economic Pathways: a machinelearning approach, EarthArxiv, 2021.[99] Vito D’Orazio, Yu Lin: Forecasting conflict in Africa with automated ma-chine learning systems , International Interactions 48 (4), 714-738, 2022.[100] Prediction of the Hotspots of Food Insecurity in Chad, Gabriele Mingoli,Marcello de Wit, Benjamin Weggelaar, Barend Spanjers.[101] Khizar Abbas, Khalid Manzoor Butt, Deyi Xu, Muhammad Ali, KhanBaz, Sanwal Hussain Kharl, Mansoor Ahmed, Measurements and deter-minants of extreme multidimensional energy poverty using machine learn-ing,Energy 251, 123977, 2022[102] Victor Gómez-Escalonilla, Marie-Louise Vogt, Elisa Destro, Moussa Is-seini, Giaime Origgi, Daira Djoret, Pedro Martinez-Santos, FrancescoHolecz, Delineation of groundwater potential zones by means of ensem-ble tree supervised classification methods in the Eastern Lake Chad basin,Geocarto International 37 (25), 8924-8951, 2022[103] Chad Abeche, Detection of Hate Speech Texts Using Machine LearningAlgorithm, Preprint 2021[104] Mahamat Saleh Adoum Sanoussi, Chen Xiaohua, George K Agordzo, Ma-hamed Lamine Guindo, Abdullah MMA Al Omari, Boukhari MahamatIssa, Detection of Hate Speech Texts Using Machine Learning Algorithm,2022 IEEE 12th Annual Computing and Communication Workshop andConference (CCWC), 0266-0273, 2022.[105] VS Subrahmanian, Chiara Pulice, James F Brown, Jacob Bonen-Clark, AMachine Learning Based Model of Boko Haram, Springer, 2021227[106] Gant Basimine Chuma, Yannick Mugumaarhahama, Jean MubalamaMond, Espoir Mukengere Bagula, Adrien Byamungu Ndeko, PrinceBaraka Lucungu, Katcho Karume, Gustave Nachigera Mushagalusa, SergeSchmitz, Gully erosion susceptibility mapping using four machine learn-ing methods in Luzinzi watershed, eastern Democratic Republic of Congo,Physics and Chemistry of the Earth, Parts A/B/C, Volume 129, 2023,103295, ISSN 1474-7065, https://doi.org/10.1016/j.pce.2022.103295.[107] Augustin Bergeron, Arnaud Fournier, John Kabeya Kabeya, GabrielTourek and Jonathan L. Weigel, Using Machine Learning to Create aProperty Tax Roll: Evidence from the City of Kananga, DR Congo, In-ternational Centre for Taxation and Development, ISBN 78-1-80470-153-9,ICTD Working Paper 176, Brighton: Institute of Development Studies,DOI:10.19088/ICTD.2023.053[108] Akeem Shola Ayinde, YU Huaming, WU Kejian, Sea Level Variabilityand Predictions Using Artificial Neural Networks and Machine LearningTechniques in the Gulf of Guinea, 2023[109] Francesco Garzon, Christian Barrientos, Rigoberto Esono Anvene, FemeEsono Mba, Alejandro Fallabrino, Angela Formia, Brendan J Godley, MaryK Gonder, Carolina Martinez Prieto, Jesus Mba Ayetebe, Kristian Met-calfe, David Montgomery, Juan Nsogo, Juan-Cruz Ondo Nze, Earl Pos-sardt, Erick Ross Salazar, Manjula Tiwari, Matthew J Witt, Spatial ecol-ogy and conservation of leatherback turtles (Dermochelys coriacea) nestingin Bioko, Equatorial Guinea, Plos one 18 (6), e0286545, 2023[110] Raquel Jimenez-Fernandez, Rocio Rodriguez Vazquez, Dolores Marin-Morales, Elena Herraiz-Soria, Marta Elena Losa-Iglesias, Ricardo Becerro-de-Bengoa-Vallejo, Inmaculada Corral-Liria, Exploring Knowledge aboutFang Traditional Medicine: An Informal Health Seeking Behaviour forMedical or Cultural Afflictions in Equatorial Guinea, Healthcare 11 (6),808, 2023.[111] Celio de Sousa, Lola Fatoyinbo, Christopher Neigh, Farrel Boucka,Vanessa Angoue, Trond Larsen, Cloud-computing and machine learningin support of country-level land cover and ecosystem extent mapping inLiberia and Gabon, PLoS One 15 (1), e0227438, 2020[112] Nina Manomba-Mbadinga, Simona Niculescu, Narimane Zaabar, Jean-Bernard Mombo, Guanyao Xie, Grand Libreville (Gabon) coastline usingmachine learning and convolutional neural network detection and auto-matic extraction of the methods, Earth Resources and Environmental Re-mote Sensing/GIS Applications XIV 12734, 1273402, 2023[113] Daniel Carcereri, Paola Rizzoli, Dino Ienco, Lorenzo Bruzzone, A DeepLearning Framework for the Estimation of Forest Height from BistaticTanDEM-X Data, IEEE Journal of Selected Topics in Applied Earth Ob-servations and Remote Sensing, 2023.228[114] Daniel Carcereri, Paola Rizzoli, Dino Ienco, Lorenzo Bruzzone, A DeepLearning Framework for the Estimation of Forest Height from BistaticTanDEM-X Data , IEEE Journal of Selected Topics in Applied Earth Ob-servations and Remote Sensing, 2023[115] Kawu, Yakubu, Okorie Agwu, and Anietie Okon. ”Oil And Gas Flow RatePrediction for African Oilfields: An Overview of Existing Models, Data andUncertainties.”[116] de Sousa, Celio, Lola Fatoyinbo, Christopher Neigh, Farrel Boucka,Vanessa Angoue, and Trond Larsen. ”Cloud-computing and machine learn-ing in support of country-level land cover and ecosystem extent mappingin Liberia and Gabon.” PLoS One 15, no. 1 (2020): e0227438.[117] Ballester-Berman, J. David, and Maria Rastoll-Gimenez. ”Sensitivityanalysis of Sentinel-1 backscatter to oil palm plantations at pluriannualscale: a case study in Gabon, Africa.” Remote Sensing 13, no. 11 (2021):2075.[118] Camilleri, L., Leila Hamza, O. Yucel, V. Rodet, and Baptiste Breton.”Lessons Learned from an ESP Development of a Naturally Fractured Car-bonate Reservoir Offshore Gabon.” In SPE Annual Technical Conferenceand Exhibition. OnePetro, 2012.[119] Thombre, Madhura. ”Deconstructing Deepfake: Tracking Legal Implica-tions and Challenges.” Issue 4 Int’l JL Mgmt. and Human. 4 (2021): 2267.[120] Jaldi, Abdessalam. Artificial Intelligence Revolution in Africa: EconomicOpportunities and Legal Challenges.” Policy Center for the New South.Policy paper (2023): 13-23.[121] Peter Fisker, Jordi Gallego-Ayala, David Malmgren Hansen, ThomasPave Sohnesen, Edmundo Murrugarra ,Guiding Social Protection TargetingThrough Satellite Data in Sao Tomé and Principe, World Bank, Washing-ton, DC, 2022[122] Cristino Mandinga Bonfim da Fonseca, The role of value chains analysis inagricultural sector. The case of pepper value chains in Sao Tomé e Principe,[123] K Osypov, M Souza, F Segovia, J Etienne, P Norlund, JDS Piedade,J Costa. ”Deep Learning Seismic Inversion: A Case Study from OffshoreAngola.” 84th EAGE Annual Conference & Exhibition 2023 (1), 1-5, 2023.[124] E Santi, MP Clarizia, D Comite, L Dente, L Guerriero, N Pierdicca. ”De-tecting fire disturbances in forests by using GNSS reflectometry and ma-chine learning: A case study in Angola.” Remote Sensing of Environment270, 112878, 2022.229[125] Armstrong Manuvakola Ezequias Ngolo, Teiji Watanabe. ”Integrating geo-graphical information systems, remote sensing, and machine learning tech-niques to monitor urban expansion: an application to Luanda, Angola.”Geo-spatial Information Science 26 (3), 446-464, 2023.[126] Xueye Wang, Gaëlle Bocksberger, Thea Lautenschläger, Manfred Finckh,Paulina Meller, Gregory E O’Malley, Vicky M Oelze. ”A bioavailablestrontium isoscape of Angola with implications for the archaeology of thetransatlantic slave trade.” Journal of Archaeological Science 154, 105775,2023.[127] Zolana R Joao. ”Road Construction Assessment Model (RC-AM) to Pre-vent Contract Overbilling in Angola.” The George Washington University,2021.[128] Laurinda Paim, Graça Barros, Ana Magro, Elsa Borges da Silva, AntónioMexia, Arlindo Lima. ”Insects and fungi in stored maize in Angola.” Julius-Kühn-Archiv, 2018.[129] Geraldo AR Ramos, Lateef Akanji. ”Application of artificial intelligencefor technical screening of enhanced oil recovery methods.” Journal of Oil,Gas and Petrochemical Sciences, 2017.[130] Geraldo AR Ramos, Lateef Akanji. ”Data analysis and neuro-fuzzy tech-nique for EOR screening: Application in Angolan oilfields.” Energies 10(7), 837, 2017.[131] Hesham Hasan, Francois D Bouchet, Maximiano Chidundo, MichaelNashaat, Joào Simào, Mario De Sà, Lidia F Ferreiras, Manuel N José. ”3DWater Saturation Estimation Using AVO Inversion Output and ArtificialNeural Network; A Case Study, Offshore Angola.” Abu Dhabi InternationalPetroleum Exhibition and Conference, D011S009R002, 2020.[132] Motlhatlhedi Gaboalapswe, ”Explore and design an artificial intelligentand data analytic software model to address domestic water usage billingcrisis in Botswana urban areas,” Botho University, 2019.[133] Tshegofatso Kgame, Huiqun Wu, Jinsong Geng, ”Knowledge and Percep-tions of Artificial Intelligence Use in Screening for Diabetic Retinopathyin Botswana: Survey for Eye Care Clinicians,” Advances in EngineeringTechnology Research 5 (1), 327-327, 2023.[134] V Sampath Kumar, J Prasad, V Lakshmi Narasimhan, S Ravi, ”Appli-cation of artificial neural networks for prediction of solar radiation forBotswana,” 2017 International Conference on Energy, Communication,Data Analytics and Soft Computing (ICECDS), 3493-3501, 2017.[135] Adewole S Oladele, Vera Vokolkova, Jerome A Egwurube, ”Pavementperformance modeling using artificial intelligence approach: a case of230Botswana district gravel road networks,” J. Eng. Appl. Sci 5 (2), 23-31,2014.[136] Dr Audrey Masizana Rajeswari Chandrasekaran, Clusteting as an artifi-cial intelligence technique in drug resistance of HIV/AIDS patients: casestudy Botswana.[137] George Anderson, Audrey N Masizana-Katongo, Dimane Mpoeleng, ”Un-derstanding the Potential of Data Mining in Botswana,” www.ajocict.net6 (1), 33, 2013.[138] Smartson P Nyoni, Thabani Nyoni, ”Botswana’s Art Program SuccessStory: Evidence from the Artificial Neural Networks,” International Re-search Journal of Innovations in Engineering and Technology 5 (3), 237,2021.[139] Kagiso Ndlovu, Nate Stein, Ruth Gaopelo, Mike Annechino, MmolokiMolwantwa, Mosadikhumo Monkge, Amy Forrestel, Victoria L Williams,”Evaluating feasibility and acceptance of a mobile clinical decision supportsystem in Botswana.”[140] Boluwaji A Akinnuwesi, Stephen G Fashoto, Andile S Metfula, Adetutu NAkinnuwesi. ”Experimental Application of Machine Learning on FinancialInclusion Data for Governance in Eswatini.” Responsible Design, Imple-mentation and Use of Information and Communication Technology: 19thIFIP WG 6.11 Conference on e-Business, e-Services, and e-Society, I3E2020, Skukuza, 2020.[141] Sibongakonke Kwanele Zungu, Qi-Xian Huang, Min-Yi Chiu, Hung-MinSun. ”Response and Surveillance System for Diarrhoea Based on a PatientSymptoms Using Machine Learning: A Study on Eswatini.” IEEE Access9, 152945-152959, 2021.[142] Stephen Gbenga Fashoto, Boluwaji A Akinnuwesi, Elliot Mbunge, AndileS Metfula. ”Financial inclusion dataset classification in Eswatini using sup-port vector machine and logistic regression.” International Journal of Busi-ness Information Systems 43 (4), 507-527, 2023.[143] Nqobile Mashwama, Thokozani Shongwe. ”Proposing the best combina-tion of Renewable Energy resources available to the country of Eswatini.”2020 International Conference on Artificial Intelligence, Big Data, Com-puting and Data Communication Systems (icABCD), 1-8, 2020.[144] Smartson P Nyoni, Thabani Nyoni, Tatenda A Chihoho. ”ForecastingDaily New Covid-19 Cases in the Kingdom Of Eswatini Using ArtificialNeural Networks.” International Research Journal of Innovations in Engi-neering and Technology 5 (3), 120, 2021.231[145] Wisdom M Dlamini, Willem AJ Nel, Jimmy Hendrick. ”Machine-learnedBayesian networks reveal the invasion dynamics of Chromolaena odoratain Eswatini.” Pan African International Conference on Science, Computingand Telecommunications Book of Proceedings, 20.[146] Vusi Wonderboy Tsabedze, Ntombikayise Nomsa Mathabela, Sanni Sham-sudeen Ademola. ”A Framework for Integrating Artificial Intelligence IntoLibrary and Information Science Curricula.” Innovative Technologies forEnhancing Knowledge Access in Academic Libraries, 233-246, 2022.[147] Stephen Gbenga Fashoto, Elliot Mbunge, Gabriel Ogunleye, Johan Vanden Burg. ”Implementation of machine learning for predicting maizecrop yields using multiple linear regression and backward elimination.”Malaysian Journal of Computing (MJoC) 6 (1), 679-697, 2021.[148] Jashim Uddin Ahmed, Raihan Sharif, Quazi Tafsirul Islam, Asma Ahmed.”Harnessing emerging digital technologies: DataNet operations in eSwa-tini.” Journal of Information Technology Teaching Cases 12 (1), 79-87,2022[149] Yego, N. K., Kasozi, J., Nkurunziza, J. (2021). A Comparative Analysis ofMachine Learning Models for the Prediction of Insurance Uptake in Kenya.Data, 6(11), 116.[150] Mulungu, K., Abro, Z. A., Muriithi, W. B., Kassie, M., Khamis, F., Ki-doido, M., Tanga, C. (2023). One size does not fit all: Heterogeneous eco-nomic impact of integrated pest management practices for mango fruit fliesin Kenya - a machine learning approach. Journal of Agricultural Economics.[151] Alharahsheh, Y. E., Abdullah, M. A. (2021). Predicting individuals men-tal health status in Kenya using machine learning methods. In 12th Inter-national Conference on Information and Communication Systems (ICICS),94-98.[152] Yego, N. K., Kasozi, J., Nkurunziza, J. (2023). Optimizing Pension Par-ticipation in Kenya through Predictive Modeling: A Comparative Analysisof Tree-Based Machine Learning Algorithms and Logistic Regression Clas-sifier. Risks, 11(4), 77.[153] Pius, A. M., Ogada, K., Mwalili, T. (2021). Supervised Machine Learn-ing Modelling of Demand for Outpatient Health-Care Services in Kenyausing Artificial Neural Networks and Regression Decision Trees. In 22ndInternational Arab Conference on Information Technology (ACIT), 1-7.[154] Shah, S. Y., Saxena, S., Rani, S. P., Nelaturi, N., Gill, S., Barr, B. T.Teltumbade, M. (2023). Prediction of postpartum hemorrhage (PPH) usingmachine learning algorithms in a Kenyan population. Frontiers in GlobalWomen’s Health, 4.232[155] Ondiek, B., Waruguru, L., Njenga, S. (2023). Recommender System forSTEM Enrolment in Universities Using Machine Learning Algorithms:Case of Kenyan Universities. International Journal of Science, Technology& Management, 4(6), 1465-1476.[156] Wilson, D. L., Coyle, J. R., Thomas, E. A. (2017). Ensemble machinelearning and forecasting can achieve 99% uptime for rural handpumps.PLoS One, 12(11), e0188808.[157] Lees, T., Tseng, G., Atzberger, C., Reece, S., Dadson, S. (2022). Deeplearning for vegetation health forecasting: a case study in Kenya. RemoteSensing, 14(3), 698.[158] Gram-Hansen, B. J., Helber, P., Varatharajan, I., Azam, F., Coca-Castro,A., Kopackova, V., Bilinski, P. (2019). Mapping informal settlements in de-veloping countries using machine learning and low resolution multi-spectraldata. In Proceedings of the 2019 AAAI/ACM Conference on AI, Ethics,and Society, 361-368.[159] Orare, E. M. (2019). Travel time prediction model for Nairobi city: anapplication of machine learning algorithms. Kca University.[160] Kochulem, E., Mwaniki, D., Mutua, F. (2023). Mass Valuation of Unim-proved Land Value Case Study: Nairobi County. Journal of GeographicInformation System, 15(1), 122-139.[161] Pius Kamando, A. M. (2023). A tree based neural network for forecastingoutpatient health-care services demand–A Case Study of medical facilitiesNairobi County, Kenya. JKUAT-COPAS.[162] Kuria, M. W. (2014). Machine learning for flood forecasting: case study:Nzoia river basin, western Kenya. University of Nairobi.[163] Onyango, K. O. (2021). Twitter sentiment analysis tool for detecting crimehotspots: a case of Nairobi, Kenya. Strathmore University.[164] Magiya, J. A. (2020). PREDICTING PACKAGE DELIVERY TIME FORMOTORCYCLES IN NAIROBI.[165] Muthoka, J. M., Salakpi, E. E., Ouko, E., Yi, Z.-F., Antonarakis, A. S.,Rowhani, P. (2021). Mapping Opuntia stricta in the Arid and Semi-AridEnvironment of Kenya Using Sentinel-2 Imagery and Ensemble MachineLearning Classifiers. Remote Sensing, 13(8), 1494.[166] Omolo, F. O. (2016). A Mobile and web based application for security in-telligence gathering-a case study of Nairobi County. Strathmore University.[167] Mbani, B. O., Odera, P. A., Kenduiywo, B. K. Crime mapping usingartificial intelligence agents in Nairobi city county, Kenya, 2020233[168] Omondi, E., Boitt, M. (2020). Modeling the Spatial Distribution of SoilHeavy Metals Using Random Forest Model—A Case Study of Nairobi andThirirka Rivers’ Confluence. Journal of Geographic Information System,12(06), 597.[169] Muchuku, A. A. (2023). Assessing Recurrent Neural Networks as a Pre-diction Tool for Quoted Stock Prices on the Nairobi Securities Exchange.University of Nairobi.[170] Mohammad Mardani, Hossein Mardani, Lorenzo De Simone, SamuelVaras, Naoki Kita, Takafumi Saito. Integration of machine learning andopen access geospatial data for land cover mapping. Remote Sensing, 11(16),1907, 2019.[171] Selebalo Joseph Ramakhanna, Botle Esther Mapeshoane, Christian ThineOmuto. Carbon sequestration potential in croplands in Lesotho. EcologicalModelling, 471, 110052, 2022.[172] Lorenzo De Simone, William Ouellette, Pietro Gennari. Operational useof EO data for national land cover official statistics in Lesotho. RemoteSensing, 14(14), 3294, 2022.[173] Freddy Bangelesa, Elhadi Adam, Jasper Knight, Inos Dhau, MarubiniRamudzuli, Thabiso MMokotjomela. Predicting soil organic carbon contentusing hyperspectral remote sensing in a degraded mountain landscape inLesotho. Applied and environmental soil science, 2020, 1-11, 2020.[174] Makhamisa Senekane, Mhlambululi Mafu, Molibeli Benedict Taele.Weather Nowcasting Using Deep Learning Techniques. Data Mining-Methods, Applications and Systems, 2021.[175] Durizzo Kathrin, Isabel Günther, Kenneth Harttgen. Using machinelearning to predict health insurance enrolment and take-up of health ser-vices.[176] Baijnath Kaushik, Akshita Sharma, Akshma Chadha, Reya Sharma. Ma-chine Learning Model for Sentiment Analysis on Mental Health Issues. 15thInternational Conference on Computer and Automation Engineering (IC-CAE), 21-25, 2023.[177] Musa Adekunle Ayanwale. Evidence from Lesotho Secondary Schools onStudents’ Intention to Engage in Artificial Intelligence Learning. IEEEAFRICON, 1-6, 2023.[178] Lereko Lefela. Formulating short-term electricity demand forecasting forLesotho, 2020.[179] Lehlohonolo L Ramokanate. The use of electronic agents in trade andcommerce: a critique of the approach under the ETEC Bill 2013. LesothoLaw Journal, 23(1), 87-115, 2015.234[180] Godfred Anakpo, Umakrishnan Kollamparambil. Artificial intelligenceand average wages in Southern Africa: A panel VAR approach. Develop-ment Southern Africa, 39(4), 575-588, 2022.[181] Tahleho Tseole, Patrick Ngulube. Improving knowledge retention in thecross-border mergers of the telecommunications industry of Lesotho. Busi-ness Information Review, 39(4), 124-135, 2022.[182] Bulemba Katende, Moniek Bresser, Mashaete Kamele, Lebohang Chere,Mosa Tlahali, Rahel Milena Erhardt, Josephine Muhairwe, Irene Ayakaka,Tracy R Glass, Morten Ruhwald, Bram van Ginneken, Keelin Murphy,Margaretha de Vos, Alain Amstutz, Mathabo Mareka, Sekhele MataboMooko, Klaus Reither, Lucia Gonzalez Fernandez. Impact of a multi-diseaseintegrated screening and diagnostic model for COVID-19, TB, and HIV inLesotho. PLOS Global Public Health, 3(8), e0001488, 2023.[183] Clément Le Ludec, Maxime Cornet, Antonio A Casilli. ”The problem withannotation. Human labour and outsourcing between France and Madagas-car.” Big Data & Society, 10 (2), 20539517231188723, 2023.[184] Fahafahantsoa Rapelanoro Rabenja. ”PASSION Project: Data Collectionin Madagascar and Guinea.” Iproceedings, 8 (1), e36903, 2022.[185] Paola Tubaro, Antonio A Casilli, Marion Coville. ”The trainer, the ver-ifier, the imitator: Three ways in which human platform workers supportartificial intelligence.” Big Data & Society, 7 (1), 2053951720919776, 2020.[186] Daniele Silvestro, Stefano Goria, Thomas Sterner, Alexandre Antonelli.”Improving biodiversity protection through artificial intelligence.” Naturesustainability, 5 (5), 415-424, 2022.[187] Sandro Valerio Silva, Tobias Andermann, Alexander Zizka, Gregor Ko-zlowski, Daniele Silvestro. ”Global estimation and mapping of the conser-vation status of tree species using artificial intelligence.” Frontiers in PlantScience, 13, 839792, 2022.[188] Harimino Andriamalala Rajaonarisoa, Théodore Totozafiny, Solonante-naina Herinjaka Rakotonasy, Adolphe Andriamanga Ratiarison. ”Charac-terization of the evolution of precipitation in Southern Madagascar usingHigh Order Fuzzy Time Series.” 2021 International Conference on Electri-cal, Computer and Energy Technologies (ICECET), 1-4, 2021.[189] Rabeniaina Anjara Davio Ulrick, Rakotovao Niry Arinavalona. ”Modellingthe onset and end dates of monsoon circulation in Northern Madagascar.”Journal of Artificial Intelligence and Systems, 5 (1), 79-90, 2023.[190] JB Koto, TR Ramahefy, S Randrianja. ”Extraction of Knowledge fromCivil Status Data (Surname and First Name) By Artificial Intelligence.235[191] Paola Tubaro, Antonio A Casilli. ”Micro-work, artificial intelligence andthe automotive industry.” Journal of Industrial and Business Economics,46, 333-345, 2019.[192] TR Rasamoela, J Szpytko. ”How to implement the use of telematics inthe transport system in Antananarivo, Madagascar?” Archives of TransportSystem Telematics, 13, 2020.[193] Manuel Dominguez-Rodrigo et al. ”Artificial intelligence provides greateraccuracy in the classification of modern and ancient bone surface modifi-cations.” Scientific Reports, 10 (1), 18862, 2020.[194] Matteo Giuliani, Marta Zaniolo, Andrea Castelletti, Guido Davoli, PaulBlock. ”Detecting the state of the climate system via artificial intelligenceto improve seasonal forecasts and inform reservoir operations.” Water Re-sources Research, 55 (11), 9133-9147, 2019.[195] Dominique Badariotti, Arnaud Banos, Vincent Laperrière, Jean-PierreMüller. ”SIMPEST: an agent-based model to simulate plague epidemics inMadagascar.” European Conference on Complex Systems, 25-29/9 2006,11 p., 2006.[196] Ala Saleh Alluhaidan. ”Artificial intelligence for public perception ofdrones as a tool for telecommunication technologies.” Computational In-telligence, 2022.[197] Juliette Soanaoly, Lucius Ramifidisoa , Ruffin Ambeondahy , Adolphe ARatiarison , Fuzzy Inference Modeling of Rain’s Beginning and Ending inCoastal Areas of South East Madagascar, American Journal of Sciencesand Engineering Research, Volume 4, Issue 5, 2021[198] Volatiana Marielle Ratianantitra, Jean Luc Razafindramintsa, ThomasMahatody, Claire Rasoamalalavao, Victor Manantsoa, Malagasy Abstrac-tive Text Summarization Using Scheduled Sampling Model, 2nd Interna-tional Conference on Industry 4.0 and Artificial Intelligence (ICIAI 2021),6-9, 2022[199] Clovis Grinand, Fety Rakotomalala, Valéry Gond, Romuald Vaudry, Mar-tial Bernoux, Ghislain Vieilledent, Estimating deforestation in tropical hu-mid and dry forests in Madagascar from 2000 to 2010 using multi-dateLandsat satellite images and the random forests classifier, Remote Sensingof Environment 139, 68-80, 2013[200] Amaa de Arsan Miriarison Tsikomia , Can a Budget Simulation Model ofDecentralized Territorial Authorities be Enhanced by an Artificial Intelli-gence Method, Risk in Contemporary Economy, 81-128, 2021[201] Victor Munoz, Lorenzo Bianco, Mark Sumka, Andrew Magee, Develop-ment of Machine Learning-based rainfall outlook models in Madagascar236[202] Smartson P Nyoni, Thabani Nyoni, Forecasting Infant Mortality Ratein Madagascar Using Artificial Neural Networks, International ResearchJournal of Innovations in Engineering and Technology 5 (3), 576, 2021[203] Kensuke Kawamura, Tomohiro Nishigaki, Andry Andriamananjara, Ho-bimiarantsoa Rakotonindrina, Yasuhiro Tsujimoto, Naoki Moritsuka,Michel Rabenarivo, Tantely Razafimbelo Using a one-dimensional convolu-tional neural network on visible and near-infrared spectroscopy to improvesoil phosphorus prediction in Madagascar, Remote Sensing 13 (8), 1519,2021.[204] Francis Chinyama Predicting household poverty with machine learningmethods: the case of Malawi, Faculty of Science, 2022[205] Shahrzad Gholami, Erwin Knippenberg, James Campbell, Daniel Andri-antsimba, Anusheel Kamle, Pavitraa Parthasarathy, Ria Sankar, CameronBirge, Juan Lavista Ferres Food security analysis and forecasting: A ma-chine learning case study in southern Malawi Data & Policy 4, e33, 2022[206] Giacomo Falchetta, Chisomo Kasamba, Simon C Parkinson Monitoringhydropower reliability in Malawi with satellite data and machine learningEnvironmental Research Letters 15 (1), 014011, 2020[207] Amelia V Taylor, Eva Mfutso-Bengo Towards a machine understandingof Malawi legal text Artificial Intelligence and Law, 1-11, 2021[208] Weldensie T Embaye, Yacob Abrehe Zereyesus, Bowen Chen Predictingthe rental value of houses in household surveys in Tanzania, Uganda, andMalawi: Evaluations of hedonic pricing and machine learning approachesPlos one 16 (2), e0244953, 2021[209] Andrew Reid Bell, Jennifer Zavaleta Cheek, Frazer Mataya, Patrick SWard Do as they did: Peer effects explain adoption of conservation agri-culture in Malawi Water 10 (1), 51, 2018[210] Lauren M Koerner, Michael A Chadwick, Emma J Tebbs, Mapping inva-sive strawberry guava (Psidium cattleianum) in tropical forests of Mauri-tius with Sentinel-2 and machine learning, International Journal of RemoteSensing, 43 (3), 841-872, 2022.[211] Jameel Ahmad Sowdagur, B Tawheeda B Rozbully-Sowdagur, Geer-ish Suddul, Road Accident Severity Prediction in Mauritius using Super-vised Machine Learning Algorithms, Faculty of Computing and Technology(FCT), University of Kelaniya, Sri Lanka, 2021.[212] Adams Begue, Venitha Kowlessur, Upasana Singh, Fawzi Mahomoodally,Sameerchand Pudaruth, Automatic recognition of medicinal plants usingmachine learning techniques, International Journal of Advanced ComputerScience and Applications, 8 (4), 166-175, 2017.237[213] A Zaynah Dhunny, Reena Hansa Seebocus, Zaheer Allam, MohammadYasser Chuttur, Muhammed Eltahan, Harsh Mehta, Flood prediction usingartificial neural networks: empirical evidence from Mauritius as a casestudy, Knowledge Engineering and Data Science, 3 (1), 1-10, 2020.[214] Manass Greedharry, Varun Seewoogobin, Nuzhah Gooda Sahib-Kaudeer,A smart mobile application for complaints in Mauritius, Information Sys-tems Design and Intelligent Applications: Proceedings of Fifth Interna-tional Conference INDIA 2018 Volume 2, 345-356, 2019.[215] Chandradeo Bokhoree, Kailash Balnac, Forecasting Demand for Electric-ity in Mauritius using Support Vector Machines for Sustainable EnergyPlanning, JOURNAL of ENVIRONMENTAL, 271, 2009.[216] Sameerchand Pudaruth, Mohamad Fawzi Mahomoodally, NoushreenKissoon, Fadil Chady, MedicPlant: A mobile application for the recogni-tion of medicinal plants from the Republic of Mauritius using deep learningin real-time, IAES International Journal of Artificial Intelligence, 10 (4),938, 2021.[217] Sameerchand Pudaruth, Predicting the price of used cars using machinelearning techniques, Int. J. Inf. Comput. Technol, 4 (7), 753-764, 2014.[218] Aleesha Mohamudally-Boolaky, Teemulsingh Luchowa, Kesseven Padachi,Applying the Support Vector Machine for Testing Pricing Inefficiency onthe Stock Exchange of Mauritius, Applied Economics and Finance, 6 (5),177-192, 2019.[219] Derinsha Canagareddy, Khuslendra Subarayadu, Visham Hurbungs, Amachine learning model to predict the performance of university students,Smart and Sustainable Engineering for Next Generation Applications: Pro-ceeding of the Second International Conference on Emerging Trends in Elec-trical, Electronic, 2019.[220] Ambareen Beebeejaun, Rajendra Parsad Gunputh, A Study of the In-fluence of Artificial Intelligence and Its Challenges: The Impact onEmployees of the Legal Sector of Mauritius, Global Business Review,09721509231193803, 2023.[221] Nuzhah Gooda Sahib, Marie Angele Justine Marianne, Baby Gobin-Rahimbux, Sentiment Analysis of Social Media Comments in Mauritius,2023 IEEE 13th Annual Computing and Communication Workshop andConference (CCWC), 0860-0865, 2023.[222] Kaviraj Gosaye, Raj Kishen Moloo, A mobile application for fruit fly iden-tification using Deep Transfer Learning: A Case Study for Mauritius, 2022International Conference for Advancement in Technology (ICONAT), 1-5,2022.238[223] Evans Jason Henri, Zahra Mungloo - Dilmohamud, A deep transfer learn-ing model for the identification of bird songs: A case study for Mauritius,2021 International Conference on Electrical, Computer, Communicationsand Mechatronics Engineering (ICECCME), 01-06, 2021.[224] Anish Appadoo, Yashna Gopaul, Sameerchand Pudaruth, FruVegy: AnAndroid App for the Automatic Identification of Fruits and Vegetables usingComputer Vision and Machine Learning, International Journal of Comput-ing and Digital Systems, 13 (1), 169-179, 2023.[225] Aphiwe Madubedube, Serena Coetzee, Victoria Rautenbach, Acontributor-focused intrinsic quality assessment of OpenStreetMap inMozambique using unsupervised machine learning, ISPRS InternationalJournal of Geo-Information 10 (3), 156, 2021[226] Lucrêncio Silvestre Macarringue, édson Luis Bolfe, Soltan Galano Du-verger, Edson Eyji Sano, Marcellus Marques Caldas, Marcos César Ferreira,Jurandir Zullo Junior, Lindon Fonseca Matias,Land Use and Land CoverClassification in the Northern Region of Mozambique Based on LandsatTime Series and Machine Learning, ISPRS International Journal of Geo-Information 12 (8), 342, 2023[227] Ruben Manhiça, Arnaldo Santos, José Cravino, The Impact of Artifi-cial Intelligence on a Learning Management System in a Higher EducationContext: A Position Paper, International Conference on Technology andInnovation in Learning, Teaching and Education, 454-460, 2022[228] Chandrasekar Vuppalapati, Food Security and Advanced Imaging Ra-diometer ML Models, Artificial Intelligence and Heuristics for EnhancedFood Security, 521-614, 2022[229] Timon Weitkamp, Poolad Karimi, Evaluating the Effect of Training DataSize and Composition on the Accuracy of Smallholder Irrigated AgricultureMapping in Mozambique Using Remote Sensing and Machine Learning,Remote Sensing 15 (12), 3017, 2023.[230] Hao Li, Benjamin Herfort, Wei Huang, Mohammed Zia, Alexander Zipf,Exploration of OpenStreetMap missing built-up areas using twitter hier-archical clustering and deep learning in Mozambique , ISPRS Journal ofPhotogrammetry and Remote Sensing 166, 41-51, 2020.[231] Sara Dominguez-Rodriguez, Miquel Serna-Pascual, Andrea Oletto, ShaunBarnabas, Peter Zuidewind, Els Dobbels, Siva Danaviah, Osee Behuhuma,Maria Grazia Lain, Paula Vaz, Sheila Fernandez-Luis, Tacilta Nhampossa,Elisa Lopez-Varela, Kennedy Otwombe, Afaaf Liberty, Avy Violari, Al-moustapha Issiaka Maiga, Paolo Rossi, Carlo Giaquinto, Louise Kuhn,Pablo Rojo, Alfredo Tagarro, Machine learning outperformed logistic re-gression classification even with limit sample size: A model to predict pe-diatric HIV mortality and clinical progression to AIDS , EPIICAL Consor-tium, Plos one 17 (10), e0276116, 2022239[232] Felermino DMA Ali, Andrew Caines, Jaimito LA Malavi, Towards a par-allel corpus of Portuguese and the Bantu language Emakhuwa of Mozam-bique, arXiv preprint arXiv:2104.05753, 2021[233] Annastasia Shipepe, Lannie Uwu-Khaeb, Emmanuel Awuni Kolog, MikkoApiola, Kauna Mufeti, Erkki Sutinen, Towards the Fourth Industrial Rev-olution in Namibia: An Undergraduate AI Course Africanized ,2021 IEEEFrontiers in Education Conference (FIE), 1-8, 2021[234] Petrus P Shifidi, Colin Stanley, Ambrose A Azeta, Machine Learning-Based Analytical Process for Predicting the Occurrence of Gender-BasedViolence , 2023 International Conference on Emerging Trends in Networksand Computer Communications (ETNCC), 1-8, 2023[235] Gabriel Tuhafeni Nhinda, Fungai Bhunu Shava, Cybersecurity Practicesof Rural Underserved Communities in Africa: A Case Study from North-ern Namibia , 2023 International Conference on Artificial Intelligence, BigData, Computing and Data Communication Systems (icABCD), 1-7, 2023[236] Andrea Santangeli, Yuxuan Chen, Mark Boorman, Sofia Sales Ligero,Guillermo Albert Garcia, Semi-automated detection of tagged animals fromcamera trap images using artificial intelligence , Ibis 164 (4), 1123-1131,2022.[237] Anton Limbo, Nalina Suresh, Set-Sakeus Ndakolute, ValerianusHashiyana, Titus Haiduwa, Martin Mabeifam Ujakpa, Smart IrrigationSystem for Crop Farmers in Namibia , Transforming the Internet of Thingsfor Next-Generation Smart Systems, 120-131, 2021.[238] Willie Schoeman, Rory Moore, Yusof Seedat, Jeff Yu-Jen Chen, Artificialintelligence: Is South Africa ready? , Gordon Institute of Business Science(GIBS), 2021.[239] Alouis Chilunjika, Kudakwashe Intauno, Sharon R Chilunjika, Artificialintelligence and public sector human resource management in South Africa:Opportunities, challenges and prospects , SA Journal of Human ResourceManagement 20, 12, 2022[240] Kavita Behara, Ernest Bhero, John Terhile Agee, Vinay Gonela, Artificialintelligence in medical diagnostics: A review from a South African context,Scientific African, e01360, 2022[241] Rudzani Mulaudzi, Ritesh Ajoodha, An exploration of machine learningmodels to forecast the unemployment rate of South Africa: a univariateapproach, 2020 2nd international multidisciplinary information technologyand engineering conference (IMITEC), 1-7, 2020[242] Eduan Kotzé, Burgert A Senekal, Walter Daelemans, Automatic classi-fication of social media reports on violent incidents in South Africa usingmachine learning, South African Journal of Science 116 (3-4), 1-8, 2020240[243] Tendani Mutavhatsindi, Caston Sigauke, Rendani Mbuvha, Forecastinghourly global horizontal solar irradiance in South Africa using machinelearning models, IEEE Access 8, 198872-198885, 2020[244] Lateef B Amusa, Annah V Bengesai, Hafiz TA Khan, Predicting the vul-nerability of women to intimate partner violence in South Africa: Evidencefrom tree-based machine learning techniques, Journal of interpersonal vio-lence 37 (7-8), NP5228-NP5245, 2022[245] Marvin Ludwig, Theunis Morgenthal, Florian Detsch, Thomas P Hig-ginbottom, Maite Lezama Valdes, Thomas Nauß, Hanna Meyer, Machinelearning and multi-sensor based modelling of woody vegetation in theMolopo Area, South Africa, Remote Sensing of Environment 222, 195-203,2019[246] Ibidun Christiana Obagbuwa, Ademola P Abidoye, South Africa crimevisualization, trends analysis, and prediction using machine learning linearregression technique, Applied Computational Intelligence and Soft Com-puting 2021, 1-14, 2021[247] Sangiwe Moyo, Tuan Nguyen Doan, Jessica Ann Yun, Ndumiso Tshuma,Application of machine learning models in predicting length of stay amonghealthcare workers in underserved communities in South Africa, Humanresources for health 16 (1), 1-9, 2018[248] Danlu Zhang, Linlin Du, Wenhao Wang, Qingyang Zhu, Jianzhao Bi,Noah Scovronick, Mogesh Naidoo, Rebecca M Garland, Yang Liu, A ma-chine learning model to estimate ambient PM2.5 concentrations in indus-trialized highveld region of South Africa, Remote sensing of environment266, 112713, 2021[249] Oluwafemi Oriola, Eduan Kotzé, Evaluating machine learning techniquesfor detecting offensive and hate speech in South African tweets, IEEE Ac-cess 8, 21496-21509, 2020[250] Tumisho Billson Mokgonyane, Tshephisho Joseph Sefara, Thipe IsaiahModipa, Mercy Mosibudi Mogale, Madimetja Jonas Manamela, PhutiJohn Manamela, Automatic speaker recognition system based on machinelearning algorithms, 2019 Southern African Universities Power EngineeringConference/Robotics and Mechatronics/Pattern Recognition Associationof South Africa (SAUPEC/RobMech/PRASA), 141-146, 2019[251] Casper Du Plessis, George Van Zijl, Johan Van Tol, Alen Manyevere,Machine learning digital soil mapping to inform gully erosion mitigationmeasures in the Eastern Cape, South Africa, Geoderma 368, 114287, 2020[252] Patrick Martineau, Swadhin K Behera, Masami Nonaka, Ratnam Jayan-thi, Takayoshi Ikeda, Noboru Minakawa, Philip Kruger, Qavanisi E241Mabunda, Predicting malaria outbreaks from sea surface temperature vari-ability up to 9 months ahead in Limpopo, South Africa, using machinelearning, Frontiers in Public Health 10, 962377, 2022[253] Irvin D Shandu, Iqra Atif, An Integration of Geospatial Modelling andMachine Learning Techniques for Mapping Groundwater Potential Zonesin Nelson Mandela Bay, South Africa , Water 15 (19), 3447, 2023[254] Festus Fatai Adedoyin, Nicholas Mavengere, Alfred Mutanga, A simula-tion experiment on ICT and patent intensity in South Africa: An applica-tion of the novel dynamic ARDL machine learning model , TechnologicalForecasting and Social Change 185, 122044, 2022[255] Daniel Page, Yudhvir Seetharam, Christo Auret, Machine learning andmanager selection: evidence from South Africa , International Journal ofEmerging Markets, 2023[256] Sulaimon Afolabi, Warrie Warrie, Oluwatobi Banjo, Opeoluwa Iwashokun,Abimbola Olawale, Naledi Ngqambela, Fata Soliu, Olawumi Olasunkanmi,Folorunso Sakinat, Sibusiso Matshika, When and where? Proactively pre-dicting traffic accident in South Africa: our machine learning competitionwinning approach , International Journal of Society Systems Science 13 (2),151-170, 2021[257] Seun O Olukanmi, Fulufhelo V Nelwamondo, Nnamdi I Nwulu, UtilizingGoogle Search Data with deep learning, machine learning and time seriesmodeling to forecast influenza-like illnesses in South Africa , IEEE Access9, 126822-126836, 2021.[258] Alexandre Lacoste, Alexandra Luccioni, Victor Schmidt, Thomas Dan-dres, Quantifying the carbon emissions of machine learning, arXiv preprintarXiv:1910.09700, 2019.[259] Selaki Ivy Ramarope, Olawale Samuel Fatoba, Tien-Chien Jen, Hydro-power generation forecast in South Africa based on Machine Learning (ML)models , Scientific African 22, e01981, 2023[260] Mohammed Majam, Mothepane Phatsoane, Keith Hanna, Charles Faul,Lovkesh Arora, Sarvesh Makthal, Akhil Kumar, Kashyap Jois, SamantaTresha Lalla-Edward, Utility of a machine-guided tool for assessing risk be-havior associated with contracting HIV in three sites in South Africa: pro-tocol for an in-field evaluation , JMIR Research Protocols 10 (12), e30304,2021[261] Rachel T Esra, Jacques Carstens, Janne Estill, Ricky Stoch, Sue LeRoux, Tonderai Mabuto, Michael Eisenstein, Olivia Keiser, Mhari Maskew,Matthew P Fox, Lucien De Voux, Kieran Sharpey-Schafer, Historical visitattendance as predictor of treatment interruption in South African HIVpatients: Extension of a validated machine learning model , PLOS GlobalPublic Health 3 (7), e0002105, 2023242[262] Rachel Esra, Jacques Carstens, Sue Le Roux, Tonderai Mabuto, MichaelEisenstein, Olivia Keiser, Erol Orel, Aziza Merzouki, Lucien De Voux,Mhari Maskew, Kieran Sharpey-Schafer, Validation and improvement ofa machine learning model to predict interruptions in antiretroviral treat-ment in South Africa , JAIDS Journal of Acquired Immune DeficiencySyndromes, 10.1097, 2022.[263] Riyad Ismail, Onisimo Mutanga, Lalit Kumar, Modeling the PotentialDistribution of Pine Forests Susceptible to Sirex Noctilio Infestations inMpumalanga, South Africa, Transactions in GIS 14 (5), 709-726, 2010[264] Rudzani Mulaudzi, Ritesh Ajoodha, Application of deep learning to fore-cast the South African unemployment rate: a multivariate approach , 2020IEEE Asia-Pacific Conference on Computer Science and Data Engineering(CSDE), 1-6, 2020[265] Kassahun Birhanu Tadesse, Megersa Olumana Dinka, Water Quality ClassModeling Using Machine Learning Algorithms at Roodeplaat Dam, SouthAfrica, Preprints, 2023.[266] Christopher Odoom, Eric Teye Mensah, Sarah Mensah Fobi, Daniel Ma-posa, Predictive Analysis of Misuse of Alcohol and Drugs using MachineLearning Algorithms: The Case of using an Imbalanced Dataset from SouthAfrica, Alexander Boateng, Appl. Math 17 (2), 261-271, 2023[267] Tshephisho Joseph Sefara, Madimetja Jonas Manamela, Promise TshepisoMalatji, Text-based language identification for some of the under-resourcedlanguages of South Africa,2016 International Conference on Advances inComputing and Communication Engineering (ICACCE), 303-307, 2016[268] Wim Naudé, Amy Bray, Celina Lee, Crowdsourcing Artificial Intelligencein Africa: Findings from a Machine Learning Contest , IZA DiscussionPaper, 2021[269] Willem Fourie, Teresa Kroesen, Jurgens de Lange, Using machine learningto improve the science-policy-society interface on the SDGs in South Africa,South African SDG Hub[270] Dossou, B. F. P., Tonja, A. L., Emezue, C. C., Olatunji, T., Etori, N.A., Osei, S., Adewumi, T. P., Singh, S. (2023). ”Adapting Pretrained ASRModels to Low-resource Clinical Speech using Epistemic Uncertainty-basedData Selection.” Preprint.[271] Dossou, B. F. P., Houndayi, I. B., Zantou, P., Hacheme, G. (2023). ”Fon-MTL: Towards Multitask Learning for the Fon Language.” In EMNLP2023.[272] Dossou, B. F. P., Gbenou, Y. K. S., Nono, M. G. (2023). ”PretrainedVision Models for Predicting High-Risk Breast Cancer Stage.” In ICLR2023.243[273] Olatunji, T., Afonja, T., Dossou, B. F. P., Tonja, A. L., Emezue, C. C.,Rufai, A. M., Singh, S. (2023). ”AfriNames: Most ASR models ’butcher’African Names.” In Interspeech 2023.[274] Liu, D., Jain, M., Dossou, B. F. P., Shen, Q., Lahlou, S., Goyal, A.,Malkin, N., Emezue, C. C., Zhang, D., Hassen, N., Ji, X., Kawaguchi, K.,Bengio, Y. (2023). ”GFlowOut: Dropout with Generative Flow Networks.”In ICML 2023.[275] Ogundepo, O., Gwadabe, T., Rivera, C., Clark, J., Ruder, S., Adelani, D.I., Dossou, B. F. P., et al. (2023). ”AfriQA: Cross-lingual Open-RetrievalQuestion Answering for African Languages.” In EMNLP 2023.[276] Olatunji, T., Afonja, T., Yadavalli, A., Emezue, C. C., Singh, S., Dos-sou, B. F. P., et al. (2023). ”AfriSpeech-200: Pan-African Accented SpeechDataset for Clinical and General Domain ASR.” In EMNLP 2023.[277] Leong, C., Shandilya, H., Dossou, B. F. P., Tonja, A. L., Mathew, J., Omo-tayo, A.-H., Yousuf, O., Akinjobi, Z., Emezue, C. C., Muhammad, S., Ko-lawole, S., Adewumi, T. P. (2023). ”Adapting to the Low-Resource Double-Bind: Investigating Low-Compute Methods on Low-Resource African Lan-guages.” In ICLR 2023.[278] Adelani, D. I., Masiak, M., Azime, I. A., Alabi, J. O., Tonja, A. L., Mwase,C., Ogundepo, O., Dossou, B. F. P., et al. (2023). ”MasakhaNEWS: NewsTopic Classification for African languages.” In AACL 2023.[279] Dione, C. M. B., Adelani, D. I., Nabende, P., Dossou, B. F. P., et al.(2023). ”MasakhaPOS: Part-of-Speech Tagging for Typologically DiverseAfrican Languages.” In ACL 2023.[280] Nigatu, H. H., Tonja, A. L., Kalita, J. (2023). ”The Less the Merrier? In-vestigating Language Representation in Multilingual Models.” In EMNLP2023.[281] Marivate, V., Mots’Oehli, M., Wagnerinst, V., Lastrucci, R., Dzingirai, I.(2023). ”PuoBERTa: Training and evaluation of a curated language modelfor Setswana.” In SACAIR 2023.[282] Mbuvha, R., Adelani, D. I., Mutavhatsindi, T., Rakhuhu, T., Mauda,A., Maumela, T. J., Masindi, A., Rananga, S., Marivate, V., Marwala, T.”MphayaNER: Named Entity Recognition for Tshivenda.”[283] Myoya, R., Banda, F., Marivate, V., Modupe, A. (2023). ”Fine tuningmultilingual pretrained african language models ” In ICLR 2023.[284] Ngomane, D., Marivate, V., Abbott, J., Mabuya, R. (2023). ”Unsuper-vised Cross-lingual Word Embedding Representation for English-isiZulu.”In RAIL 2023.244[285] Siminyu, K., Abbott, J., Tubosun, K., Mthembu, A. T., Ramkilowan,A., Oladimeji, B., Cell Press Publication. ”Consultative engagement ofstakeholders toward a roadmap for African language technologies.”[286] Ojo, J., Ogueji, K., Stenetorp, P., Adelani, D. I.. How good are LargeLanguage Models on African Languages?[287] Obvious Nchimunya Chilyabanyama, Roma Chilengi, Michelo Simuyandi,Caroline C Chisenga, Masuzyo Chirwa, Kalongo Hamusonde, Rakesh Ku-mar Saroj, Najeeha Talat Iqbal, Innocent Ngaruye, Samuel Bosomprah,Performance of machine learning classifiers in classifying stunting amongunder-five children in Zambia , Children 9 (7), 1082, 2022[288] Dhruv Grover, Sebastian Bauhoff, Jed Friedman, Using supervised learn-ing to select audit targets in performance-based financing in health: Anexample from Zambia , PloS one 14 (1), e0211262, 2019[289] Sipiwe Chihana, Jameson Mbale, Nchimunya Chaamwe, Leveraging Ma-chine Learning for Ambient Air Pollutant Prediction: The Zambian MiningEnvironment Context , Proceedings of International Conference for ICT(ICICT)-Zambia 4 (1), 1-5, 2022[290] Varun Kshirsagar, Jerzy Wieczorek, Sharada Ramanathan, Rachel Wells,Household poverty classification in data-scarce environments: A machinelearning approach, arXiv preprint arXiv:1711.06813, 2017[291] Kwabena Adu-Ababio, Aliisa Koivisto, Eliya Lungu, Evaristo Mwale,Jonathan Msoni, Kangwa Musole, Estimating tax gaps in Zambia, Preprint2023.[292] Sahar Kazemzadeh, Jin Yu, Shahar Jamshy, Rory Pilgrim, Zaid Nabulsi,Christina Chen, Neeral Beladia, Charles Lau, Scott Mayer McKinney, ThadHughes, Atilla P Kiraly, Sreenivasa Raju Kalidindi, Monde Muyoyeta,Jameson Malemela, Ting Shih, Greg S Corrado, Lily Peng, Katherine Chou,Po-Hsuan Cameron Chen, Yun Liu, Krish Eswaran, Daniel Tse, ShravyaShetty, Shruthi Prabhakara, Deep learning detection of active pulmonarytuberculosis at chest radiography matched the clinical performance of ra-diologists, Radiology 306 (1), 124-137, 2023[293] Valentina Bellemo, Zhan Wei Lim, Gilbert Lim, Quang Nguyen, MichelleYT Yip, Yuchen Xie, Xin Qi Lee, Haslina Hamzah, Jinyi Ho, Gavin SiewWei Tan, Wynne Hsu, Mong Li Lee, Sobha Sivaprasad, Geeta Menon,Tien Yin Wong, Daniel SW Ting, Application of Deep Learning System toScreen for Diabetic Retinopathy in an Underprivileged African Populationwith Diabetes, Investigative Ophthalmology & Visual Science 60 (9), 1439-1439, 2019.[294] John Batani, Manoj Sewak Maharaj, Towards data-driven pediatrics inZimbabwe, 2022 International Conference on Artificial Intelligence, BigData, Computing and Data Communication Systems (icABCD), 1-7, 2022245[295] Vehicle Damage model classification for Zimbabwe Insurance Sector us-ing MobileNetV2 and DenseNet121, Pavlov Takudzwa Mpinyuri, EdmoreTarambiwa, 2022 1st Zimbabwe Conference of Information and Communi-cation Technologies (ZCICT), 1-5, 2022[296] Panashe Chiurunge, Agripah Kandiero, Framework for Prediction of De-pression Among Adolescents Using Machine Learning: A Case of Zim-babwe, Integrating Digital Health Strategies for Effective Administration,310-344, 2023.[297] Daxelle Sakubu, Kelly Joelle Gatore Sinigirira, David Niyukuri, Predict-ing malaria dynamics in Burundi using deep Learning Models , Preprint2023.[298] CIAT, Banana pest risk assessment along banana trade axes running fromlow to high altitude sites, in the Eastern DR Congo and in Burundi, TheAlliance of Bioversity International and the International Center for Trop-ical Agriculture (CIAT), 2020.[299] Shafika Isaacs, Florence Ngombo , ICT in Education in the Comoros,Population, 2005[300] Harriet Joseph Ottenheimer: Spelling Shinzwani: Dictionary construc-tion and orthographic choice in the Comoro Islands, Written Language &Literacy 4 (1), 15-29, 2001[301] Abdi, Hamda and Mohamed, Assabo and Omar, Abdou Idris, Sky Tem-perature Forecasting in Djibouti: An Integrated Approach Using MeasuredClimate Data and Artificial Neural Networks. Preprint SSRN, 2023.[302] Denis Pastory Rubanga, Sergio Azael May-Cuevas, Yessy Arve-lyna, and Sawahiko Shimada. (2023). Fracture-fault detection us-ing deep learning with stepwise elimination from satellite imagesin Djibouti. Geomate Journal, 25(108), 241 - 248. Retrieved fromhttps://geomatejournal.com/geomate/article/view/3967[303] Shawn B. Hood, Matthew J. Cracknell, Michael F. Gazley, Anya M. Read-ing, Improved supervised classification of bedrock in areas of transportedoverburden: Applying domain expertise at Kerkasha, Eritrea, AppliedComputing and Geosciences, 2019, 100001, ISSN 2590-1974,[304] Michael B. Richman, Lance M. Leslie, Zewdu T. Segele, ClassifyingDrought in Ethiopia Using Machine Learning, Procedia Computer Science,Volume 95, 2016, Pages 229-236, ISSN 1877-0509,[305] Eshetu, K.D., Alamirew, T. and Woldesenbet, T.A. Interpretable machinelearning for predicting evaporation from Awash reservoirs, Ethiopia. EarthSci Inform (2023). https://doi.org/10.1007/s12145-023-01063-y246[306] Martin Kuradusenge, Santhi Kumaran, Marco Zennaro, ”Rainfall-inducedlandslide prediction using machine learning models: The case of NgororeroDistrict, Rwanda,” International journal of environmental research andpublic health 17 (11), 4147, 2020[307] Roger Muremyi, Dominique Haughton, Ignace Kabano, François Niragire:”Prediction of out-of-pocket health expenditures in Rwanda using machinelearning techniques,” Pan African Medical Journal 37 (1), 2020[308] Yves Uwiragiye, Mbezele Junior Yannick Ngaba, Mengzhen Zhao, AhmedS Elrys, Gerard BM Heuvelink, Jianbin Zhou, ”Modelling and mappingsoil nutrient depletion in humid highlands of East Africa using ensemblemachine learning: A case study from Rwanda,” Catena 217, 106499, 2022[309] Francois Mbonyinshuti, Joseph Nkurunziza, Japhet Niyobuhungiro, EgideKayitare, ”The prediction of essential medicines demand: a machine learn-ing approach using consumption data in Rwanda,” Processes 10 (1), 26,2021[310] Similien Ndagijimana, Ignace Habimana Kabano, Emmanuel Masabo,Jean Marie Ntaganda, ”Prediction of Stunting Among Under-5 Childrenin Rwanda Using Machine Learning Techniques,” Journal of PreventiveMedicine and Public Health 56 (1), 41, 2023[311] Emmanuel Masabo, Joseph Nzabanita, Innocent Ngaruye, Charles Ru-ranga, Jean Paul Nizeyimana, Jeanne Uwonkunda, Denis Ndanguza, ”Earlydetection of students at risk of poor performance in Rwanda higher educa-tion using machine learning techniques,” International Journal of Informa-tion Technology, 1-10, 2023[312] Theogene Rizinde, Innocent Ngaruye, Nathan D Cahill, ”Comparing Ma-chine Learning Classifiers for Predicting Hospital Readmission of HeartFailure Patients in Rwanda,” Journal of Personalized Medicine 13 (9), 1393,2023[313] Jean Baptiste Nsengiyumva, Roberto Valentino, ”Predicting landslide sus-ceptibility and risks using GIS-based machine learning simulations, case ofupper Nyabarongo catchment,” Geomatics, Natural Hazards and Risk 11(1), 1250-1277, 2020[314] Paul Brimble, Patrick McSharry, Felix Bachofer, Jonathan Bower, An-dreas Braun, ”Using machine learning and remote sensing to value propertyin Rwanda,” International Growth Centre, 2020[315] Arthur Kwizera, Niranjan Kissoon, Ndidiamaka Musa, Olivier Urayeneza,Pierre Mujyarugamba, Andrew J Patterson, Lori Harmon, Joseph CFarmer, Martin W Dünser, Jens Meier, ”A machine Learning-Based triagetool for children with acute infection in a low resource setting,” PediatricCritical Care Medicine 20 (12), e524-e530, 2019247[316] Janvier Niyitegeka, Didacienne Mukanyiligira, Said Rutabayiro Ngoga,Emmanuel Masabo, Louis Sibomana, Raymond Ndacyayisaba, ”Employingmachine learning and internet of things for malaria outbreak prediction inRwanda,” International Workshop Soft Computing Applications, 338-364,2020[317] Martin Kuradusenge, Eric Hitimana, Damien Hanyurwimfura, PlacideRukundo, Kambombo Mtonga, Angelique Mukasine, Claudette Uwitonze,Jackson Ngabonziza, Angelique Uwamahoro, ”Crop yield prediction usingmachine learning models: case of Irish potato and maize,” Agriculture 13(1), 225, 2023[318] Gatesi Jean de Dieu, Shuai Bin, Wencheng Huang, Ntakiyemungu Math-ieu, ”Comparison of supervised machine learning algorithms for road trafficcrash prediction models in Rwanda,” Proceedings of the Institution of CivilEngineers-Transport, 1-37, 2023[319] Joseph Habiyaremye, Marco Zennaro, Chomora Mikeka, EmmanuelMasabo, Kayalvizhi Jayavel, Santhi Kumaran, ”A data-driven predictivemachine learning model for efficiently storing temperature-sensitive medicalproducts, such as vaccines: Case study: Pharmacies in Rwanda,” Journalof Healthcare Engineering 2021, 1-12, 2021[320] Leopord Uwamahoro, Papias Niyigena, ”Deep Learning in GreenhouseGases Emissions from Agriculture Activities in Rwanda Using Long ShortTerm Memory Recurrent Neural Network,” Int. Res. J. Eng. Technol 6 (10),928-931, 2019[321] Ines Emera, Markon Sandor, ”Creation of farmers’ awareness on fall army-worms pest detection at early stage in Rwanda using deep learning,” 20198th International Congress on Advanced Applied Informatics (IIAI-AAI),538-541, 2019[322] Regina Kilwenge, Julius Adewopo, Zhanli Sun, Marc Schut, ”UAV-basedmapping of banana land area for village-level decision-support in Rwanda,”Remote Sensing 13 (24), 4985, 2021[323] Emily Murphy, Swathi Samuel, Joseph Cho, William Adorno, MarcelDurieux, Donald Brown, Christian Ndaribitse, ”Checkbox detection onRwandan perioperative flowsheets using convolutional neural network,”2021 Systems and Information Engineering Design Symposium (SIEDS),1-6, 2021[324] Fidelis L Bologo, Barry Rawn, Tunmise Raji, ”Ground Truths to SupportRemote-Sensing Inference of Irrigation Benefits and Effects in Rwanda,”2023 IEEE AFRICON, 1-6, 2023248[325] Francis Kipkogei, Ignace H Kabano, Belle Fille Murorunkwere, NzabanitaJoseph, ”Business success prediction in Rwanda: a comparison of tree-based models and logistic regression classifiers,” SN Business & Economics1, 1-19, 2021[326] Omar H Kombo, Santhi Kumaran, Emmanuel Ndashimye, AlastairBovim, ”An Ensemble Mode Decomposition Combined with SVR-RFModel for Prediction of Groundwater Level: The Case of Eastern RwandanAquifers,” Computer Science On-line Conference, 312-328, 2022[327] Meron Alemnew Kifle, ”IoT and ML-Based precision agriculture: A caseof Rwanda coffee,” College of science and Technology, 2022[328] Esaie Dufitimana, Theophile Niyonzima, ”Leveraging the Potential ofConvolutional Neural Network and Satellite Images to Map Informal Set-tlements in Urban Settings of the City of Kigali, Rwanda,” Rwanda Journalof Engineering, Science, Technology and Environment 5 (1), 2023[329] DJB Swanborn, Nico Fassbender, VAI Huvenne, SJ Pittman, PV Ste-fanoudis, LC Woodall, ”Geomorphological drivers of deeper reef habitatsaround Seychelles,” Coral Reefs 41 (4), 1001-1016, 2022[330] Benjamin Lee, Lucy Martin, Dimosthenis Traganos, Sylvanna Antat,Stacy K Baez, Annabelle Cupidon, Annike Faure, Jérôme Harlay, MatthewMorgan, Jeanne A Mortimer, Peter Reinartz, Gwilym Rowlands, ”Map-ping the National Seagrass Extent in Seychelles Using PlanetScope NICFIData,” Remote Sensing 15 (18), 4500, 2023[331] Daniel R Schrider, Julien Ayroles, Daniel R Matute, Andrew D Kern,”Supervised machine learning reveals introgressed loci in the genomes ofDrosophila simulans and D. sechellia,” PLoS genetics 14 (4), e1007341,2018[332] Luo Xiao, Sally W Thurston, David Ruppert, Tanzy MT Love, PhilipW Davidson, ”Bayesian models for multiple outcomes in domains withapplication to the Seychelles Child Development Study,” Journal of theAmerican Statistical Association 109 (505), 1-10, 2014[333] Abdullahi Ahmed Abdirahman, Abdirahman Osman Hashi, Ubaid Mo-hamed Dahir, Mohamed Abdirahman Elmi,Octavio Ernest Romo Ro-driguez, Comparative Analysis of Machine Learning and Deep LearningModels for Sentiment Analysis in Somali Language, SSRG InternationalJournal of Electrical and Electronics Engineering Volume 10 Issue 7, 41-52,July 2023[334] Tohid Atashbar, How Nations Become Fragile: An AI-Augmented Bird’s-Eye View (with a Case Study of South Sudan), WPIEA2023167, August11, 2023.249[335] Felix Sukums, Deogratias Mzurikwao, Deodatus Sabas, Rebecca Chaula,Juliana Mbuke, Twaha Kabika, John Kaswija, Bernard Ngowi, JosefNoll, Andrea S. Winkler, Sarah Wamala Andersson, The use of artificialintelligence-based innovations in the health sector in Tanzania: A scopingreview, Health Policy and Technology, Volume 12, Issue 1, 2023, 100728,ISSN 2211-8837, https://doi.org/10.1016/j.hlpt.2023.100728.[336] Nick Jones, Vivien Deparday, Nuala Margaret Coman, Mapping Dar esSalaam’s trees through artificial intelligence , MAY 18, 2022[337] Andrew Katumba, Joyce Nakatumba-Nabende, Jenifer Winfred Na-muyanja, Jeremy Tusubira Francis, Claire Babirye, Jonathan Mukiibi,Chodrine Mutebi, Hewiit Tusiime, Building NLP Text and Speech Datasetsfor Low Resourced Languages in East Africa, 2022.[338] Emily Aiken, Suzanne Bellue, Dean Karlan, Chris Udry andJoshua E. Blumenstock, Machine learning and phone data can im-prove targeting of humanitarian aid. Nature 603, 864 - 870 (2022).https://doi.org/10.1038/s41586-022-04484-9[339] Kohnert, Dirk. ”Machine ethics and African identities: Perspectives ofartificial intelligence in Africa.” Available at SSRN 4163096 (2022).[340] Fousseni, Arafat, Mawugno Koffi Kodjo, and Assiongbon Adanlete Ad-janoh. ”Static and dynamic evaluation of wind potential in the Kara re-gion of Togo using artificial neural networks.” TH Wildau Engineering andNatural Sciences Proceedings 1 (2021).[341] Agbodjan, Marc Marin, Joseph Mung’atu, and Anthony Wanjoya. ”Pre-diction of Maize Price using Hidden Markov Chain Model: An Applicationon Grand market of Lome (TOGO).[342] Aiken, Emily, Suzanne Bellue, Joshua Blumenstock, Dean Karlan, andChristopher R. Udry. ”Estimating Impact with Surveys versus DigitalTraces: Evidence from Randomized Cash Transfers in Togo.” (2023).[343] Yomo, Mawulolo, Euloge Nicaise Yalo, Masamaéya Dadja-Toyou Gna-zou, Stephen Silliman, Isaac Larbi, and Khaldoon A. Mourad. ”Forecastingland use and land cover dynamics using combined remote sensing, machinelearning algorithm and local perception in the Agoènyivé Plateau, Togo.”Remote Sensing Applications: Society and Environment 30 (2023): 100928.[344] Dotche, Koffi Agbeblewu, Adekunlé Akim Salami, Koffi Mawugno Kodjo,Yawa Pamela CD Blu, and Yao Essemu Julien Diabo. ”Evaluating Solar En-ergy Harvesting using Artificial Neural Networks: A Case study in Togo.”In 2019 II International Conference on High Technology for SustainableDevelopment (HiTech), pp. 1-5. IEEE, 2019.250[345] Zhuangfang NaNa Yi, Naroa Zurutuza, Do-Hyung Kim, Ruben LopezMendoza, Martha Morrissey, Chuck Daniels,Nick Ingalls, Jeevan Farias,Karito Tenorio, Pilar Serrano,Sajjad Anwar, Building on our success map-ping 23,100 unmapped schools in eight countries , Development Seed, 2021[346] Sarr, Alioune Badara, and Benjamin Sultan. ”Predicting crop yields inSenegal using machine learning methods.” International Journal of Clima-tology 43, no. 4 (2023): 1817-1838.[347] Nyasulu, Chimango, Awa Diattara, Assitan Traore, Abdoulaye Deme,and Cheikh Ba. ”Towards Resilient Agriculture to Hostile Climate Changein the Sahel Region: A Case Study of Machine Learning-Based WeatherPrediction in Senegal.” Agriculture 12, no. 9 (2022): 1473.[348] MMbengue, F., Gayane Faye, Kharouna Talla, M. Adama Sarr, AndréFerrari, and Modou Mbaye. ”Evaluation Of Machine Learning Classifica-tion Methods For Rice Detection Using Earth Observation Data: Case OfSenegal.” European Scientific Journal, ESJ 18, no. 17 (2022): 214.[349] Bayet, Théophile, Timothée Brochier, Christophe Cambier, Alassane Bah,Christophe Denis, and N. Thiam JD Zucker. ”A Machine Learning ap-proach to improve the monitoring of Sustainable Development Goals: acase study in Senegalese artisanal fisheries.” CNIA (2021): 30.[350] Dia, Yoro, Lamine Faty, Aba Diop, Ousmane Sall, and Tony Tona Landu.”A Hybrid Model for Predicting Road Accident Severity in Senegal.” InInternational Conference on Artificial Intelligence Science and Applications,pp. 1-12. Cham: Springer Nature Switzerland, 2022.[351] State, Laura, Hadrien Salat, Stefania Rubrichi, and Zbigniew Smoreda.”Explainability in Practice: Estimating Electrification Rates from MobilePhone Data in Senegal.” In World Conference on Explainable ArtificialIntelligence, pp. 110-125. Cham: Springer Nature Switzerland, 2023.[352] Sarkodie, Samuel Asumadu, Emmanuel Ackom, Festus Victor Bekun, andPhebe Asantewaa Owusu. ”Energy - climate - economy - population nexus:an empirical analysis in Kenya, Senegal, and Eswatini.” Sustainability 12,no. 15 (2020): 6202.[353] Kebe, M. A. M., Maodo Sene, and Nafissatou Diagne. ”An experienceof detection and classification of Quality-Of-Service problems in MV/LVdistribution substations using artificial intelligence: Senegal case study.”(2023): 1719-1723.[354] Seck, Djamal Abdoul Nasser, and Fatima Bintou Rassoul Diakité. ”Su-pervised Machine Learning Models for the Prediction of renal failure inSenegal.” In 2023 International Conference on Control, Artificial Intelli-gence, Robotics & Optimization (ICCAIRO), pp. 94-98. IEEE, 2023.251[355] Lee, Kamwoo, Jeanine Braithwaite, and Michel Atchikpa. ”Understandingcolonial legacy and environmental issues in Senegal through language use.”In Social, Cultural, and Behavioral Modeling: 13th International Confer-ence, SBP-BRiMS 2020, Washington, DC, USA, October 18 - 21, 2020,Proceedings 13, pp. 23-34. Springer International Publishing, 2020.[356] Dione, Cheikh M. Bamba, Jonas Kuhn, and Sina Zarrieß. ”Design andDevelopment of Part-of-Speech-Tagging Resources for Wolof (Niger-Congo,spoken in Senegal).” In LREC. 2010.[357] Alla, Lo, Dione Cheikh Bamba, Nguer Elhadji Mamadou, Ba Sileye O.Ba, and Lo Moussa. ”Using lstm to translate french to senegalese locallanguages: Wolof as a case study.” arXiv preprint arXiv:2004.13840 (2020).[358] Dia, Yoro, Lamine Faty, Ousmane Sall, Tony Tona Landu, and EdouardNgor Sarr. ”Predicting the Severity of the Road Accidents in Senegal: AnEmpirical Study.” In 2022 the 5th International Conference on Data Stor-age and Data Engineering, pp. 91-96. 2022.[359] Dione, Doudou, Idy Diop, Ibrahima Gueye, Bassirou Ngom, and Sidi Mo-hamed Farssi. ”Proposal for an IoT-based e-health model in developingcountries: case of Senegal.” In 2021 International Conference on Electrical,Computer and Energy Technologies (ICECET), pp. 1-7. IEEE, 2021.[360] Özdogan, Mutlu, and Ajit Govind. ”Three decades of forest cover changein Senegal from remote sensing.” (2022).[361] Traoré, Fousseini, Suwadu Sakho Jimbira, and Leysa Maty Sall. ”Non-linear price transmission in the rice market in Senegal: a model-basedrecursive partitioning approach.” Applied Economics 54, no. 20 (2022):2343-2355.[362] Drame, M., Seck, D.A.N. and Ndiaye, B.S., 2023, August. Analysis andForecast of Energy Demand in Senegal with a SARIMA Model and anLSTM Neural Network. In The International Conference on Deep Learning,Big Data and Blockchain (pp. 129-140). Cham: Springer Nature Switzer-land.[363] Sarron, Julien, Cheikh Amet Bassirou Sane, Philippe Borianne, EricMalézieux, Thibault Nordey, Frédéric Normand, Paterne Diatta, YougaNiang, and Emile Faye. ”Is machine learning efficient for mango yield esti-mation when used under heterogeneous field conditions?.” In XXX Inter-national Horticultural Congress IHC2018: VII Conference on Landscapeand Urban Horticulture, IV Conference on 1279, pp. 201-208. 2018.[364] Diop, Papa Moussa, Jin Takamoto, Yuji Nakamura, and Morikazu Naka-mura. ”A machine learning approach to classification of Okra.” In 202035th International Technical Conference on Circuits/Systems, Computersand Communications (ITC-CSCC), pp. 254-257. IEEE, 2020.252[365] Moustapha Mbaye, Mouhamadou, Fadi MH Abu Salem, and Nadia Tahiri.”A new machine learning workflow to create an optimal waiting list in hos-pitals.” In Proceedings of the 2023 7th International Conference on Medicaland Health Informatics, pp. 159-163. 2023[366] Oyebode, Oladapo, and Rita Orji. ”Detecting factors responsible for di-abetes prevalence in Nigeria using social media and machine learning.” In2019 15th International Conference on Network and Service Management(CNSM), pp. 1-4. IEEE, 2019.[367] Obite, Chukwudi Paul, Angela Chukwu, Desmond ChekwubeBartholomew, Ugochinyere Ihuoma Nwosu, and Gladys EzenwanyiEsiaba. ”Classical and machine learning modeling of crude oil productionin Nigeria: Identification of an eminent model for application.” EnergyReports 7 (2021): 3497-3505.[368] Ighile, Eseosa Halima, Hiroaki Shirakawa, and Hiroki Tanikawa. ”Appli-cation of GIS and machine learning to predict flood areas in Nigeria.”Sustainability 14, no. 9 (2022): 5039.[369] Villacis, Alexis H., Syed Badruddoza, Ashok K. Mishra, and Joaquin May-orga. ”The role of recall periods when predicting food insecurity: A machinelearning application in Nigeria.” Global Food Security 36 (2023): 100671.[370] McKenzie, David J., and Dario Sansone. ”Man vs. machine in predict-ing successful entrepreneurs: evidence from a business plan competition inNigeria.” (2017).[371] Goni, Ibrahim, and Murtala Mohammad. ”Machine learning approach tomobile forensics framework for cyber crime detection in Nigeria.” Journalof Computer Science Research 2, no. 4 (2020): 1-6.[372] Folorunso, Sakinat O., Sulaimon A. Afolabi, and Adeoye B. Owodeyi.”Dissecting the genre of Nigerian music with machine learning models.”Journal of King Saud University-Computer and Information Sciences 34,no. 8 (2022): 6266-6279.[373] Lawal, Zaharaddeen Karami, Hayati Yassin, and Rufai Yusuf Zakari.”Flood prediction using machine learning models: a case study of Kebbistate Nigeria.” In 2021 IEEE Asia-Pacific Conference on Computer Scienceand Data Engineering (CSDE), pp. 1-6. IEEE, 2021.[374] Nwankwo, Mmesoma Peace, Ndukaku Macdonald Onyeizu, EmmanuelChibuogu Asogwa, Chukwuogo Okwuchukwu Ejike, and Okechukwu J. Ob-ulezi. ”Prediction of House Prices in Lagos-Nigeria Using Machine LearningModels.” European Journal of Theoretical and Applied Sciences 1, no. 5(2023): 313-326.253[375] Panjala, Pranay, Murali Krishna Gumma, Hakeem Ayinde Ajeigbe, Mur-tala Muhammad Badamasi, Kumara Charyulu Deevi, and Ramadjita Tabo.”Identifying Suitable Watersheds across Nigeria Using Biophysical Param-eters and Machine Learning Algorithms for Agri - Planning.” ISPRS Inter-national Journal of Geo-Information 11, no. 8 (2022): 416.[376] McKenzie, David, and Dario Sansone. ”Predicting entrepreneurial successis hard: Evidence from a business plan competition in Nigeria.” Journal ofDevelopment Economics 141 (2019): 102369.[377] Gladys, Ajibolu Oyinkansola, and Akinola Solomon Olalekan. ”A machinelearning model for predicting colour trends in the textile fashion industryin south-west Nigeria.” International Journal on Integrated Education 4,no. 2 (2021): 174-188.[378] Odeniyi, Olufemi A., Mabel E. Adeosun, and Tayo P. Ogundunmade.”Prediction of terrorist activities in Nigeria using machine learning mod-els.” Innovations 71 (2022): 87-96.[379] Ogundunmade, T. P., and A. A. Adepoju. ”Modelling Liquefied PetroleumGas Prices in Nigeria Using Time Series Machine Learning Models.” ModEcon Manag 1, no. 5 (2022): 3.[380] Ekubo, Ebiemi Allen, and Bukohwo Michael Esiefarienrhe. ”Using ma-chine learning to predict low academic performance at a Nigerian univer-sity.” The African Journal of Information and Communication 2022, no. 30(2022): 1-33.[381] Ekubo, Ebiemi Allen, and Bukohwo Michael Esiefarienrhe. ”Using ma-chine learning to predict low academic performance at a Nigerian univer-sity.” The African Journal of Information and Communication 2022, no. 30(2022): 1-33.[382] Muhammad, Bilyaminu, and Asaf Varol. ”A Symptom-Based MachineLearning Model for Malaria Diagnosis in Nigeria.” In 2021 9th InternationalSymposium on Digital Forensics and Security (ISDFS), pp. 1-6. IEEE, 2021.[383] Salele, Badiya, Yakubu Aminu Dodo, Dalhatu Aliyu Sani, Mo-hammed Awad Abuhussain, Barno Sayfutdinovna Abdullaeva, and AdamBrysiewicz. ”Run-off modelling of pervious and impervious areas using cou-ple SWAT and a novel machine learning model in cross-rivers state Nigeria.”Water Science & Technology 88, no. 7 (2023): 1893-1909.[384] Adeeyo, Y. A., and A. A. Osinaike. ”Modeling the Oil Viscosity of NigerianCrudes Using Machine Learning.” In SPE Nigeria Annual InternationalConference and Exhibition, p. D021S002R004. SPE, 2023.[385] Jean, Neal, Marshall Burke, Michael Xie, W. Matthew Davis, David B.Lobell, and Stefano Ermon. ”Combining satellite imagery and machinelearning to predict poverty.” Science 353, no. 6301 (2016): 790-794.254[386] Ibrahim, Esther Shupel, Claas Nendel, Bahareh Kamali, Efron NdukeGajere, and Patrick Hostert. ”Predicting potato diseases in smallholderagricultural areas of Nigeria using machine learning and remote sensing-based climate data.” PhytoFrontiers ja (2023).[387] Oyebode, Oladapo, and Rita Orji. ”Social media and sentiment analy-sis: the Nigeria presidential election 2019.” In 2019 IEEE 10th Annual In-formation Technology, Electronics and Mobile Communication Conference(IEMCON), pp. 0140-0146. IEEE, 2019.[388] Eneanya, Obiora A., Jorge Cano, Ilaria Dorigatti, Ifeoma Anagbogu,Chukwu Okoronkwo, Tini Garske, and Christl A. Donnelly. ”Environmen-tal suitability for lymphatic filariasis in Nigeria.” Parasites & vectors 11,no. 1 (2018): 1-13.[389] Obulezi, Okechukwu J., Eberechukwu Q. Chinedu, Dorathy O. Oramulu,Harrison O. Etaga, Ndukaku M. Onyeizu, and Chukwuogo OkwuchukwuEjike. ”Machine Learning models for predicting transportation costs in-flated by fuel subsidy removal policy in Nigeria.” International ResearchJournal of Modernization in Engineering Technology and Science 5: 1053-1070.[390] Mbaoma, Oliver C., Akinyemi O. Ogunkeyede, Adedoyin A. Adebayo, andPrekeyi Tawari-Fufeyin. ”Geospatial and Machine Learning Driven Air Pol-lution Modelling Using Vehicular Load and Climatic Variables in Agbarho,Delta State, Nigeria.” European Journal of Environment and Earth Sci-ences 3, no. 3 (2022): 34-41.[391] Achara, Miriam, Emeka J. Okereke, Nwulu Stephen Onyemere, andUfuoma Earnest Ofierohor. ”Financial Institution Readiness and Adoptionof Machine Learning Algorithm and Performance of Select Banks in RiversState, Nigeria.” Asian Journal of Economics, Finance and Management(2023): 180-192.[392] Oyewola, David O., Asabe Ibrahim, Joshua A. Kwanamu, and EmmanuelGbenga Dada. ”A new auditory algorithm in stock market prediction onoil and gas sector in Nigerian stock exchange.” Soft computing letters 3(2021): 100013.[393] Odeniyi, Olufemi A., Mabel E. Adeosun, and Tayo P. Ogundunmade.”Prediction of terrorist activities in Nigeria using machine learning mod-els.” Innovations 71 (2022): 87-96.[394] Ogundunmade, T. P., and A. A. Adepoju. ”Modelling Liquefied PetroleumGas Prices in Nigeria Using Time Series Machine Learning Models.” ModEcon Manag 1, no. 5 (2022): 3.255[395] Ekubo, Ebiemi Allen, and Bukohwo Michael Esiefarienrhe. ”Using ma-chine learning to predict low academic performance at a Nigerian univer-sity.” The African Journal of Information and Communication 2022, no. 30(2022): 1-33.[396] Muhammad, Bilyaminu, and Asaf Varol. ”A Symptom-Based MachineLearning Model for Malaria Diagnosis in Nigeria.” In 2021 9th InternationalSymposium on Digital Forensics and Security (ISDFS), pp. 1-6. IEEE, 2021.[397] Hamidou Mamane Moumouni, Noma Talibi Soumäıla, Boureima Seibou,Adekunlé Akim Salami, Attoumane Kosso Moustapha, and MadougouSäıdou. ”Electrical Charge of Niamey City Modelisation by Neural Net-work.” Science 7, no. 1 (2019): 13-19.[398] Schulz, Dario, He Yin, Bernhard Tischbein, Sarah Verleysdonk, RabaniAdamou, and Navneet Kumar. ”Land use mapping using Sentinel-1 andSentinel-2 time series in a heterogeneous landscape in Niger, Sahel.” ISPRSJournal of Photogrammetry and Remote Sensing 178 (2021): 97-111.[399] Painter, Thomas M. ”Adult Literacy and Cooperative Training Programsinthe Niamey Department Development Project (Niger Republic): Evalu-ation of Phase I Activities and Recommendations for Phase II.” (1980).[400] Eli, Taleb and Sidi Hamou, Lalla Aisha. ”Investigating the Factors ThatInfluence Students’ Choice of English Studies As a Major: the Case ofUniversity of Nouakchott Al Aasriya, Mauritania.” International Journalof Technology, Innovation and Management (IJTIM), 2(1), 2022.[401] Aragón, Diego Gómez. ”Identifying Desert Locust Breeding Areas byMeans of Earth Observation in Mauritania.” PhD diss., Universidad deValladolid, 2019.[402] El Arby Chrif, Mohamed El Moustapha, Saleck, Moustapha Mohamed,N’Diaye, Aichetou Cheikh Mohamedou, and Mahmoud, El Benany Mo-hamed. ”Business Intelligence Models for e-Government in Mauritania: ASurvey.” In The International Conference on Artificial Intelligence andSmart Environment, pp. 307-312. Cham: Springer International Publish-ing, 2022.[403] Abeidi, Hassine Ali, Kerim, Bedine, and Nanne, Mohamedade Farouk.”A Study for Remote Monitoring of Water Points in Mauritania Based onIoT (LoRa) Technology.” Indonesian Journal of Electrical Engineering andInformatics (IJEEI), 10(2), 2022, pp. 398-409.[404] Gómez-Escalonilla, Victor, Pedro Martinez-Santos, and Miguel Martin-Loeches. ”Preprocessing approaches in machine-learning-based groundwa-ter potential mapping: an application to the Koulikoro and Bamako re-gions, Mali.” Hydrology and Earth System Sciences, 26(2), 2022, pp. 221-243.256[405] Gómez-Escalonilla, Victor, Oumou Diancoumba, D. Y. Traoré, E. Mon-tero, M. Martin-Loeches, and P. Martinez-Santos. ”Multiclass spatial pre-dictions of borehole yield in southern Mali by means of machine learningclassifiers.” Journal of Hydrology: Regional Studies, 44, 2022, 101245.[406] Boudinaud, Laure, and Sacha Alex Orenstein. ”Assessing Cropland Aban-donment from Violent Conflict in Central Mali with SENTINEL-2 andGoogle Earth Engine.” The International Archives of the Photogrammetry,Remote Sensing and Spatial Information Sciences, 46, 2021, pp. 9-15.[407] Traore, Boukaye Boubacar, Bernard Kamsu-Foguem, and Fana Tangara.”Deep convolution neural network for image recognition.” Ecological infor-matics, 48, 2018, pp. 257-268.[408] Rodriguez-Gonzalez, Alejandro, Massimiliano Zanin, and ErnestinaMenasalvas-Ruiz. ”Public health and epidemiology informatics: can artifi-cial intelligence help future global challenges? An overview of antimicrobialresistance and impact of climate change in disease epidemiology.” Yearbookof Medical Informatics, 28(01), 2019, pp. 224-231.[409] Rajanarayanan, S., Lea Sorilla Nisperos, and J. R. Basal. ”Recognitionof pathogens using image classification based on improved recurrent neuralnetwork with LSTM.” ICTACT Journal on Soft Computing, 9(2), 2019.[410] Bizimana, Jean-Claude, David A. Bessler, and Francis Keita. ”Threedecades of market liberalization policy in Mali: An analysis of the grainmarkets.” No. 333-2016-14672, 2016.[411] Briggs, Morgan. ”SatDash: An Interactive Dashboard for Assessing LandDamage in Nigeria and Mali.” In ACM SIGCAS Conference on Computingand Sustainable Societies, 2021, pp. 100-114.[412] Hirblinger, A.T., Wählisch, M., Keator, K., McNaboe, C., Duursma, A.,Karlsrud, J., Sticher, V., Verjee, A., Kyselova, T., Kwaja, C.M. and Per-era, S. ”Making peace with uncertainty: Reflections on the role of digitaltechnology in peace processes beyond the data hype.” 2023.[413] Diallo, Abdoul Karim, and Kheireddine Abainia. ”Offensive LanguageDetection in Code-Mixed Bambara-French Corpus: Evaluating machinelearning and deep learning classifiers.” In 2023 International Conferenceon Decision Aid Sciences and Applications (DASA), IEEE, 2023, pp. 121-125.[414] Olkers, Tim, Shuang Liu, and Oliver Musshoff. ”A typology of Malianfarmers and their credit repayment performance-An unsupervised machinelearning approach.” 2023.[415] Traoré, Mohamed, and Hamadoun Bokar. ”Principal Component Analysisof Hydrochemical Data From Groundwater Resources in Bamako.”257[416] de Sousa, Celio, Lola Fatoyinbo, Christopher Neigh, Farrel Boucka,Vanessa Angoue, and Trond Larsen. ”Cloud-computing and machine learn-ing in support of country-level land cover and ecosystem extent mappingin Liberia and Gabon.” PLoS One, 15(1), 2020, e0227438.[417] Blair, Robert A., Christopher Blattman, and Alexandra Hartman. ”Pre-dicting local violence: Evidence from a panel survey in Liberia.” Journalof Peace Research, 54(2), 2017, pp. 298-312.[418] Burch Jones, Evelyn J. ”Remote Sensing, Machine Learning, and ChangeDetection Applications for Land Cover Studies: A Geospatial Analysis ofTwo Liberian Proposed Protected Areas.” 2021.[419] Yazdani, Mehrdad, Mai H. Nguyen, Jessica Block, Daniel Crawl, NaroaZurutuza, Dohyung Kim, Gordon Hanson, and Ilkay Altintas. ”Scalabledetection of rural schools in Africa using convolutional neural networksand satellite imagery.” In 2018 IEEE/ACM International Conference onUtility and Cloud Computing Companion (UCC Companion), IEEE, 2018,pp. 142-147.[420] Miller, E., Luo, D., Deparday, V., Giovando, C., Gevaert, C.M., Emanuele,R., Kassel, S., Needham, E., Lipstein, G., Bull, P. and Slavitt, I. ”An openmachine learning challenge to map urban development and resilience indiverse African cities from aerial imagery.” In AGU Fall Meeting 2019,December 2019.[421] Carreiras, Joao MB, Maria J. Vasconcelos, and Richard M. Lucas. ”Under-standing the relationship between aboveground biomass and ALOS PAL-SAR data in the forests of Guinea-Bissau (West Africa).” Remote Sensingof Environment, 121, 2012, pp. 426-442.[422] Pereira, Sofia C., Catarina Lopes, and Joao Pedro Pedroso. ”MappingCashew Orchards in Cantanhez National Park (Guinea-Bissau).” RemoteSensing Applications: Society and Environment, 26, 2022, 100746.[423] Martiarena, Merlin Leunda, and Marina Padrao Temudo. ”Endogenouslearning and innovation in African smallholder agriculture: lessons fromGuinea-Bissau.” The Journal of Agricultural Education and Extension,2023, pp. 1-19.[424] Badji, Aliou, Youssou Dieng, Ibrahima Diop, Papa Alioune Cisse, andBoubacar Diouf. ”Automatic Speaker Recognition (ASR) Application inthe monitoring of PLHIV in the cross-border area between the Gambia,Guinea-Bissau and Senegal.” In Proceedings of the 10th International Con-ference on Information Systems and Technologies, 2020, pp. 1-9.[425] Yehadji, Dègninou, Geraldine Gray, Carlos Arias Vicente, PetrosIsaakidis, Abdourahimi Diallo, Saa André Kamano, and Thierno SaidouDiallo. ”Development of machine learning algorithms to predict viral loadsuppression among HIV patients in Conakry (Guinea).” (2023).258[426] Colubri, A., Silver, T., Fradet, T., Retzepi, K., Fry, B. and Sabeti,P., 2016. Transforming clinical data into actionable prognosis models:machine-learning framework and field-deployable app to predict outcomeof Ebola patients. PLoS neglected tropical diseases, 10(3), p.e0004549.[427] Fatima Ezzahra SALAMATE, Mohamed EL AZHARI, Jamal Zahi: Ma-chine learning in epidemiology: Characterization of risk factors relatedto the occurrence of pulmonary and extra pulmonary tuberculosis in theprovince of Settat, International Journal of Computer Engineering andData Science (IJCEDS) 2 (3), 2022[428] Nyamekye, Clement, Samuel Kwofie, Benjamin Ghansah, EmmanuelAgyapong, and Linda Appiah Boamah. ”Assessing urban growth in Ghanausing machine learning and intensity analysis: A case study of the NewJuaben Municipality.” Land Use Policy, 99, 2020, 105057.[429] Abdul Muhsin Zambang, M., Jiang, H., and Wahab, L. ”Modeling vehicleownership with machine learning techniques in the Greater Tamale Area,Ghana.” PloS one, 16(2), 2021, p.e0246044.[430] Andoh, John, Louis Asiedu, Anani Lotsi, and Charlotte Chapman-Wardy.”Statistical analysis of public sentiment on the Ghanaian government: amachine learning approach.” Advances in Human-Computer Interaction,2021, pp. 1-7.[431] Twumasi, Clement, and Juliet Twumasi. ”Machine learning algorithmsfor forecasting and backcasting blood demand data with missing valuesand outliers: A study of Tema General Hospital of Ghana.” InternationalJournal of Forecasting, 38(3), 2022, pp. 1258-1277.[432] Owusu-kumih, Eric, and Ebenezer Asare Aboagye. ”Historical Analysisand Time Series Forecasting of Internet Data Usage and Revenues in GhanaUsing Machine Learning Based Facebook Prophet Model.” Available atSSRN 4290081, 2022.[433] Wahab, Lukuman, and Haobin Jiang. ”Severity prediction of motorcyclecrashes with machine learning methods.” International journal of crash-worthiness, 25(5), 2020, pp. 485-492.[434] Guenther, Martin. ”Local effects of artisanal mining: Empirical ev-idence from Ghana.” In Presentation at the International Confer-ence’Environmental Economics: A Focus on Natural Resources’, Orleans,2018.[435] Adu, Kwabena, Patrick Kwabena Mensah, Mighty Abra Ayidzoe, ObedAppiah, Ebenezer Quayson, Christopher Bombie Ninfaakang, and MichaelOpoku. ”GC3558: An open-source annotated dataset of Ghana currencyimages for classification modeling.” Data in Brief, 45, 2022, 108616.259[436] Agbodza, Paul A. ”Ito calculus-machine learning projection of forward USdollar-Ghana cedi rates.” In 2019 International Conference on Computing,Computational Modelling and Applications (ICCMA), IEEE, 2019, pp. 102-1022.[437] Barenblitt, Abigail, Amanda Payton, David Lagomasino, Temilola Fatoy-inbo, Kofi Asare, Kenneth Aidoo, Hugo Pigott et al. ”Spatial distribu-tion of artisanal goldmining in Ghana: Using machine learning and GoogleEarth Engine to quantify conversion of vegetation to gold mines.” AuthoreaPreprints, 2022.[438] Agbenyega, Diana Ayorkor, John Andoh, Samuel Iddi, and Louis Asiedu.”Modelling Customs Revenue in Ghana Using Novel Time Series Methods.”Applied Computational Intelligence and Soft Computing, 2022.[439] Agbodza, Paul A. : ”Machine Learning of Jump Dynamics in US Dollar-Ghana Cedi Exchange Returns.” In 2019 International Conference on Com-puter, Data Science and Applications (ICDSA), IEEE, 2019, pp. 1-6.[440] Signé, Landry, and Chelsea Johnson. Africa’s Mining Potential: Trends,Opportunities, Challenges and Strategies. (2021).[441] Gallwey, Jane, Carlo Robiati, John Coggan, Declan Vogt, and MatthewEyre. ”A Sentinel-2 based multispectral convolutional neural network fordetecting artisanal small-scale mining in Ghana: Applying deep learningto shallow mining.” Remote Sensing of Environment 248 (2020): 111970.[442] Kansake, Bruno Ayaga, Felix Adaania Kaba, Nelson Kofi Dumakor-Dupey, and Clement Kweku Arthur. ”The future of mining in Ghana: Arestakeholders prepared for the adoption of autonomous mining systems?.”Resources Policy 63 (2019): 101411.[443] Forson, Eric Dominic, and Prince Ofori Amponsah. ”Mineral prospectivitymapping over the Gomoa Area of Ghana’s southern Kibi-Winneba beltusing support vector machine and naive bayes.” Journal of African EarthSciences 206 (2023): 105024.[444] Arhinful, Kofi Owusu, and George Agyei. ”Fuel Consumption Predictionin Shovel-Truck System of Surface Mine Using Artificial Neural Network.”Journal of Science and Technology Research 5, no. 2 (2023).[445] Issaka, Yakubu, and Bernard Kumi-Boateng. ”Artificial intelligence tech-niques for predicting tidal effects based on geographic locations in Ghana.”Geodesy and Cartography 46, no. 1 (2020): 1-7.[446] Elvis Korku Avenyo, Learning to innovate in urban informal enterprisesof Ghana: Role of apprenticeship and formal interaction, Scientific African,Volume 20, 2023, e01740, ISSN 2468-2276,260[447] John Bosco Dramani, Prince Boakye Frimpong, Kwame Ansere Ofori-Mensah, Modelling the informal sector and energy consumption in Ghana,Social Sciences & Humanities Open, Volume 6, Issue 1, 2022, 100354, ISSN2590-2911,[448] Franic, J. What do we really know about the drivers of undeclared work?An evaluation of the current state of affairs using machine learning. AI &Soc (2022). https://doi.org/10.1007/s00146-022-01490[449] Mohammed, Nuredin I., Alexander Jarde, Grant Mackenzie, UmbertoD’Alessandro, and David Jeffries. ”Deploying Machine Learning Models Us-ing Progressive Web Applications: Implementation Using a Neural NetworkPrediction Model for Pneumonia Related Child Mortality in The Gambia.”Frontiers in Public Health, 9 (2022): 772620.[450] Ingram, Will, Leen Saeb, Seyed Sadr, Rob Hygate, and Fayyaz Ali Memon.”A preliminary investigation of smart rural water distribution systems inthe Gambia.” (2018).[451] Ferraris, Stefano, Rosa Meo, Stefano Pinardi, Matteo Salis, and GabrieleSartor. ”Machine Learning as a Strategic Tool for Helping Cocoa Farmersin Côte D’Ivoire.” Sensors 23, no. 17 (2023): 7632.[452] Eneanya, O.A., Koudou, B.G., Aboulaye, M., Elvis, A.A., Souleymane,Y., Kouakou, M.M., Weil, G.J. and Fischer, P.U., 2021. Progress towardsonchocerciasis elimination in Côte d’Ivoire: A geospatial modelling study.PLoS neglected tropical diseases, 15(2), p.e0009091.[453] Gong, Xianda, Heike Wex, Thomas Müller, Silvia Henning, JensVoigtländer, Alfred Wiedensohler, and Frank Stratmann. ”Understandingaerosol microphysical properties from 10 years of data collected at CaboVerde based on an unsupervised machine learning classification.” Atmo-spheric Chemistry and Physics, 22, no. 8 (2022): 5175-5194.[454] Varela, Danilson, Maria M. Romeiras, and Luis Silva. ”Implications ofclimate change on the distribution and conservation of Cabo Verde endemictrees.” Global Ecology and Conservation, 34 (2022): e02025.[455] Guzman-Vilca, Wilmer Cristobal, Manuel Castillo-Cara, and RodrigoM. Carrillo-Larco. ”Development, validation, and application of a ma-chine learning model to estimate salt consumption in 54 countries.” Elife,11,(2022): e72930.[456] Tiengo, Rafaela, José MR Pacheco, Jéssica Garcia Uchôa, and Artur JoséFreire Gil. ”Using Sentinel-1 GRD SAR data for volcanic eruptions moni-toring: the case-study of Fogo Volcano (Cabo Verde) in 2014/2015.” Inter-continental Geoinformation Days, (2021): 22-25.261[457] Harvey, David, Wessel Valkenburg, and Amara Amara. ”Predictingmalaria epidemics in Burkina Faso with machine learning”. PLoS One,16, no. 6 (2021): e0253302.[458] Schug, Franz, Akpona Okujeni, Janine Hauer, Patrick Hostert, Jonas O.Nielsen, and Sebastian van der Linden. ”Mapping patterns of urban devel-opment in Ouagadougou, Burkina Faso, using machine learning regressionmodeling with bi-seasonal Landsat time series.” Remote Sensing of Envi-ronment, 210 (2018): 217-228.[459] Yameogo, Baowendsomme Armel, Tounwendyam Frédéric Ouedraogo,and Constantin Zongo. On the Machine Learning Models to Predict Town-Scale Energy Consumption in Burkina Faso. In International Conferenceon Emerging Technologies for Developing Countries, pp. 65-73. Cham:Springer Nature Switzerland, 2022.[460] Bournas, Nasreddine, Abdoulaye Touré, Mahamoud Balboné, Palalé SamaZagré, Abdoulaye Ouédraogo, Kanita Khaled, Alexander Prikhodko, andJean Legault, Use of machine learning techniques on airborne geophysicaldata for mineral resources exploration in Burkina Faso, ASEG ExtendedAbstracts, 2019, no. 1 (2019), 1-4.[461] Zongo, Constantin, Frederic T Ouedraogo, and M. A. L. O. Sadouanouan,Forecasting models for monthly energy consumption using machine learningin Burkina Faso, In 2023 IEEE AFRICON, pp. 1-3. IEEE, 2023.[462] Hounkpatin, Kpade OL, Aymar Y. Bossa, Yacouba Yira, Moüınou A.Igue, and Brice A. Sinsin, Assessment of the soil fertility status in Benin(West Africa) - Digital soil mapping using machine learning, GeodermaRegional, 28 (2022): e00444.[463] Selvaraj, M.G., Vergara, A., Montenegro, F., Ruiz, H.A., Safari, N., Ray-maekers, D., Ocimati, W., Ntamwira, J., Tits, L., Omondi, A.B. andBlomme, G.: Detection of banana plants and their major diseases throughaerial images and machine learning methods: A case study in DR Congoand Republic of Benin, ISPRS Journal of Photogrammetry and RemoteSensing,169, pp.110-124, 2020.[464] Guenoupkati, Agbassou, Adekunlé Akim Salami, Mawugno Koffi Kodjo,and Kossi Napo. Short-Term Electricity Generation Forecasting Using Ma-chine Learning Algorithms: A Case Study of the Benin Electricity Com-munity (CEB), TH Wildau Engineering and Natural Sciences Proceedings,1,(2021).[465] Agossou, C., Atchadé, M.N., Djibril, A.M. and Kurisheva, S.V., Mathe-matical modeling and machine learning for public health decision-making:the case of breast cancer in Benin, Mathematical Biosciences and Engineer-ing, 19(2), pp.1697-1720, 2022262[466] Akpoti, Komlavi, Amos T. Kabo-Bah, Elliott R. Dossou-Yovo, ThomasA. Groen, and Sander J. Zwart. Mapping suitability for rice productionin inland valley landscapes in Benin and Togo using environmental nichemodeling, Science of the total environment, 709 (2020), 136-165.263

Clean Full Text

Language

Doi

Arxiv

Mag

Acl

Pmid

Pmcid

Pub Date

Pub Year

Journal Name

Journal Volume

Journal Page

Publication Types

Tldr

Tldr Version

Generated Tldr

Search Term Used

Reference Count

Citation Count

Influential Citation Count

Last Update

Status

Aws Job

Last Checked

Modified

Created