Research

It is generally accepted that the deployment of AI in Africa will generate new employment opportunities (through new types of jobs), accelerate organisational efficiency (through automated processes and decision-making) and improve public service delivery (through more responsive and personalised attention) in well-prepared settings (Pillay, 2018). However, existing infrastructure and skills gaps inhibit the ability of African countries to leverage the potential of AI, in terms of scientific contributions to AI development, local production of AI-based goods and services, and use of AI to deliver public services. For example, in the context of an already unequal society in which employment opportunities tend to be less readily accessible to people in resource-constrained environments, the efficiency-related benefits of integrating AI into existing systems will also bypass these populations. Furthermore, as is occurring in other parts of the world, the introduction of AI is likely to lead to critical challenges such as misuse or uncritical acceptance of biased automated decision-making (Gwagwa Arthur et al., 2020; Trajtenberg, 2018). The ability to face these challenges will largely be influenced by the extent to which nations have nurtured an appropriately skilled workforce, not only with technical skills but also with the knowledge and training to govern the deployment of AI. However, in spite of a growing community of Africa-based AI enthusiasts (Snow, 2019), developing a home-grown talent base to execute this vision remains a challenge due to infrastructural and educational system limitations. A recent review of AI capacity in Africa (Butcher et al., 2021) identified numerous capacity challenges including lack of AI experts and lecturers; limited capacity of educational institutions; poor funding for AI research, infrastructure and entrepreneurship; and male dominance in the AI community. Diversity challenges are also evident, especially gender-related but also in terms of other groups such as people with disabilities (Butcher et al., 2021; Centre for Intellectual Property and Information Technology Law, n.d.). Despite the recognition of a skills gap in Africa, knowledge about the actual state of AI skills across the continent is quite thin, as is knowledge about institutions and initiatives working to fill the gap (apart from a few notably well-publicised programmes such as the African Masters of Machine Intelligence in Rwanda). Few measures of capacity exist, with current tools often depending on proxy indicators (e.g., Government AI Readiness index) or having limited data on Africa (e.g., the AI Index, AI Talent Report). Equally unclear is the extent to which and how AI technology itself is being utilised within programmes aimed at boosting employment or building workforce capacity across economic sectors. The two case studies presented in this report contribute to narrowing the knowledge gap on AI skills capacity building and AI use in workforce development in Africa, especially as it relates to the public sector. The Africa AI Accelerator provides a case of AI as an output of capacity-building activities (developing a workforce able to create, deploy and govern AI products and systems), while the Harambee Youth Accelerator is a case study of AI as an input to development agendas (using AI to improve general labour force participation or educational outcomes). The cases address skills and capacity challenges from the perspectives of entrepreneurship and employment – two critical components of most national economic development plans – and formulated with some degree of collaboration between the public and private sector

Artificial Intelligence (AI) has become a main feature of news coverage and social media discourse. News and social media coverage can drive the ongoing discussions about the use of AI and influence attitudes towards it. The study used mixed methodology (automatic content analysis and manual coding) to establish the framing of AI on Twitter in Kenya and South Africa. The analysis mainly focused on determining the different local and regional narratives in tweets and retweets in the countries of study pertaining to AI in different categories. The study substantiated the claims, and general views, espoused in the analyzed tweets with data from local and international resources to determine their veracity. A total of 256 tweets from Kenya and 516 tweets from South Africa pertaining to AI sent between 2016 – 2021 were analyzed. These tweets were categorized into 7 different groups: (i) automation and job replacement, (ii) education, (iii) AI and development, (iv) commercial services, (v) health, (vi) AI and governance, and (vii) ethics and regulation, and then further delineated according to 3 sentiments: positive, negative or neutral tweet. The sentiments conveyed by the compiled tweets across these 7 categories was assessed. Study findings showed that, in general, there is still a tendency toward an optimistic view of the possible impact of AI on solving problems in Kenya and South Africa. The differences in negative and positive sentiments across the different categories skews, for the most part, toward higher positive sentiments in Kenya on a particular topic than in South Africa. Finally, the sentiments, both positive and negative, espoused in these tweet mirror those of Global North countries concerning AI, even when the on-the-ground-realities do not support these concerns.

The study identified sex disaggregation in terms of overall disaggregation, industry and position (managerial or otherwise). The total number of men and women is 977 and 406, respectively in the Artificial Intelligence (AI) workforce in Africa as revealed in a survey of 160 companies across 21 countries. The project maps the gender composition of AI projects and companies across Africa, capturing the diversity struggles particular to AI start-ups. It examines what those struggles exemplify in an African context and determines the mechanisms that can be put in place to curb them.

Development and adoption of artificial intelligence (AI) in Africa has occurred slowly relative to developed countries. A vibrant AI ecosystem is growing on the continent. Due to the unique geographical, cultural and political nature of the continent, the 4th industrial revolution on the continent is evolving differently from its global counterparts. The motivations for development of AI systems, the parties involved, and the impact of the AI ecosystem on the continent are therefore best analyzed and framed through a unique African lens. This paper seeks to begin this process by developing a conceptual framework to characterize the parties involved in the African AI ecosystem. i.e., the African AI stakeholder. Identification of these stakeholders will aid in determining their interests, responsibilities and accountability and will provide a basis for the development and implementation of an equitable AI ecosystem. It is our goal that this framework, ultimately, be used to guide the contributions from the African AI perspective in global dialogues on ethics, bias, inclusion and similar topics in the AI sphere.