Africa and Human-Centric AI

By the year 2050, the population in Africa will increase by 63% from roughly 1.5 billion to 2.5 billion. This means 25% of the world’s population will be African. Due to a number of challenges, Africa and many nations in the Global South lag significantly behind nations in the Global North in the AI revolution. Yet, with the youngest and poorest population, Africa stands to gain more than any other region from an AI revolution in education.

In centering the transformative role of AI in education, the Think7 Italy 2024 Science and Digitalization task force policy brief, Towards Safe, Secure, and Trustworthy AI: Implementing the G7 AI Hiroshima Policy Framework highlights the need for a human-centric approach to the rapid development of algorithmic and machine capability. We will need a multidisciplinary approach to education if we want to maintain the human advantage over AI systems that have the capacity to create AI algorithms.

Future AI developers must not be trained merely as scientists or engineers, but also be well-educated in the history and importance of human culture and society. This dual knowledge of the humanities, as well as the design, development, and deployment of AI is essential to the implementation of AI systems capable of fostering human connection.

The African Union’s efforts to collaborate with tech companies and work with other countries to establish governance standards offers a unique chance to promote human-centric AI educational models that can unlock human potential and transform economies.

Collaboration, Infrastructure and Norms

When the African Union launched the “Continental Artificial Intelligence Strategy” this year, its aim was to strengthen AI infrastructure and governance capabilities as prerequisites to achieving ethical regulation and the innovation much of Africa desperately needs. To promote a unified Africa-centric AI approach that lifts all of Africa out of its impoverished colonial past, the African Union seeks to bolster a business investment-friendly environment that enhances educational partnerships with tech companies like OpenAI, Google, Meta, Microsoft, TikTok and WorldCoin. At the same time, it receives ethical and technical guidance from major world organizations like UNESCO and the German International Development Agency (Deutsche Gesellschaft für Internationale Zusammenarbeit or GIZ). Nations like Kenya and Ghana are models of how to become a hub of Information and Communications Technology (ICT) infrastructure investment and AI employment and educational opportunity.

Strategic discussions between the BRICS nations and the African Union focus on overcoming development deficits in areas of energy, digital infrastructure and education. At the same time, they seek to establish governance frameworks, e.g. Agile Governance models, to promote ethical and responsible AI development. The 2023 formation of the BRICS Institute of Future Networks in Johannesburg offers Africa the partnership it needs to 1) further expand cooperation on AI, 2) promote information exchange and technological cooperation, 3) fend off risks, and 4) develop AI governance frameworks and standards to make AI technologies more secure, reliable, controllable and equitable.

Inequalities Stifle Progress

However, an Africa-BRICS partnership with tech companies to implement AI-driven educational solutions will confront several significant challenges. First, there is the digital infrastructure disparity between the nations. The 2023 study by Access Partnership and Google, AI in Africa: Unlocking Potential, Igniting Progress, indicates that adoption of AI tools by businesses in just four countries — Nigeria, Ghana, Kenya and South Africa — could net up to $136 billion for the African economy. However promising, this data belies the fact that many countries in Africa continue to grapple with limited internet connectivity, poor electrical infrastructure and inadequate access to modern devices. This problematic gap in infrastructure poses a significant barrier not only to business development, but also to the implementation of AI-powered educational tools. According to an International Telecommunication Union (ITU) report, internet penetration in sub-Saharan Africa was roughly 28% in 2019, compared to over 80% in most BRICS nations. Collaboration between the African Union, BRICS and tech companies requires a concerted effort to upgrade digital infrastructures, particularly in rural and underserved areas, and to make investments in more reliable internet service, electricity and technological devices. Coupled with ethical considerations surrounding data privacy and security in AI education systems (see Towards Safe, Secure, and Trustworthy AI policy brief), addressing the digital infrastructure divide is a fundamental prerequisite to establish meaningful collaboration that serves the educational interests of all learners.

Shifting Investments to Unlock Potential

In addition to the digital infrastructure divide and data privacy/security concerns, the geographic concentration of the IT talent pool poses challenges to the development of AI systems in Africa. Simply put, the computer information specialists, data scientists and engineers who create and maintain oversight of AI algorithm development are not located in Africa. Although China produces close to half of the world’s top AI researchers, the United States is where the top AI experts prefer to work. As a result, 60% of global AI research is done in the U.S. and 50% is done by foreign born workers.

Tech companies can support the AI transformation by working with local educators, policymakers, and cultural experts to help fund large-scale initiatives in AI education led by AI experts willing to relocate to Africa.

These experts can provide training in STEM fields to unlock the potential of a population that is 40% digitally and non-digitally literate. STEM training and increased digital literacy would exponentially increase the ability of Africa’s growing population to access and effectively use AI educational tools.

Closely connected to the geographic concentration of AI experts in the United States is the lack of actual computing resources in Africa. The UN Development Programme found that adequate computational power for research and development was available to only five percent of data scientists — or 550 persons in a cohort of 11,000 — when it analyzed computer usage data across Zindi, the largest network of AI talent in Africa. Seventy-five percent of the 550 persons were constrained by significant access limitations to GPU computer chips capable of handling the complex mathematical calculations to train AI models. The implications are clear. Africa does not have the critical resources it needs to effectively innovate across priority sectors as outlined in the Italy-Africa Mattei Plan.

Development of the AI research talent pool and better resources are needed to close the education and technology resource gaps that prevent citizen access to a host of otherwise productive AI tools. In order to meet all of the aforementioned challenges, the African Union, tech companies and other global institutions must prioritize a shift from the current global demand for AI systems performance — with its focus on scalability and efficiency — toward a more equitable and sustainable future that balances AI access for all stakeholders, including the continent of Africa and the rest of the Global South.

Dr. Paul A. McAllister, founder and president of Global Leaders in Unity and Evolvement (GLUE), is co-author of the policy brief Towards Safe, Secure, and Trustworthy AI: Implementing the G7 AI Hiroshima Policy Framework, and a member of the Think7 Italy Task Force on Science and Digitalization.

Featured image designed by Freepik