Key Takeaways
- African Union ministers met in Tangier to discuss AI as countries rush to adopt strategies, attract investment, and build digital infrastructure.
- The policy focus is shifting from mere AI adoption to questions of ownership, governance, and the terms under which AI systems are deployed.
- Nigeria, Kenya, Egypt, and Ghana have launched national AI strategies; Ghana calls AI a “sovereign capability,” and 49 AU members endorsed the Africa Declaration on Artificial Intelligence.
- Policy formulation faces practical hurdles, exemplified by South Africa’s withdrawn draft AI policy after AI‑generated, unverified content was discovered.
- Global AI industry fragmentation may give African states greater bargaining leverage, as seen in regulatory pushback against projects like Starlink.
- Africa’s data‑centre capacity is under 1 % of the world’s total despite 18 % of the global population, hampered by electricity shortages.
- Kenya’s contested $1 bn Microsoft‑G42 data‑centre deal highlights tensions over energy demands, financing, and long‑term strategic dependence.
- Experts stress that partnerships must be evaluated for developmental returns, noting water‑intensive data centres and historical patterns of subsidised foreign investment.
- Concerns extend to data sovereignty, surveillance, and public trust; citizen inclusion is vital to avoid a trust gap that could hinder fintech, e‑government, and e‑commerce adoption.
- Initiatives such as the UNDP’s timbuktoo programme aim to build local capacity, ensuring Africa participates as both a contributor and a consumer in the global AI ecosystem.
Context of the African Union AI Discussion
In April, African Union ministers convened in Tangier, Morocco, to deliberate on artificial intelligence at a time when governments across the continent are hurriedly crafting national AI strategies, courting foreign investment, and expanding digital infrastructure. Beneath the optimism lay a fundamental dilemma: as overseas technology firms pour resources into data centres, cloud services, and AI platforms on African soil, how much ultimate control will African states retain over the very infrastructure that powers these technologies? The meeting highlighted a shift from mere adoption questions to deeper concerns about ownership, governance, and the conditions under which AI systems are developed and deployed.
Shift from Adoption to Ownership and Governance
For years, policy discourse centred on how governments, businesses, and public services could harness AI. Increasingly, the conversation is turning to who owns the technology, how it is governed, and what terms accompany its deployment. Several African nations have begun framing AI as a matter of sovereignty, seeking to build local capacity and lessen reliance on external providers. This reorientation reflects a broader recognition that the benefits of AI will be limited unless African countries can influence the foundational layers—hardware, data, and expertise—upon which the technology rests.
National Strategies and Continental Declarations
Nigeria, Kenya, Egypt, and Ghana have all released national AI strategies in recent years that emphasize indigenous talent development and reduced dependence on foreign tech firms. Ghana’s April‑launched strategy explicitly labels AI a “sovereign capability.” Complementing these efforts, forty‑nine African countries, together with the African Union, have endorsed the Africa Declaration on Artificial Intelligence. The declaration calls for heightened investment in African AI infrastructure, education, and innovation, and proposes coordinated financing mechanisms to strengthen the continent’s position in the global AI ecosystem.
Challenges in Policy Formulation
Translating ambition into concrete policy has proven difficult. In South Africa, a draft national AI policy was withdrawn earlier this year after officials discovered that certain passages could not be verified and appeared to have been generated by AI tools themselves. The incident underscored the practical hurdles governments face when trying to regulate a rapidly evolving field, highlighting the need for rigorous vetting processes and the risks of relying on unverified AI‑generated content during policymaking.
Global Competition and Leverage for African States
The discussion unfolds amid intensifying global rivalry over AI, as major technology companies, cloud providers, and governments vie for data, computing power, and new markets. Priyal Singh, a geopolitical analyst at Signal Risk, argued that the fragmented nature of the worldwide AI industry could actually bolster African negotiating power. He cited regulatory pushback against Starlink’s expansion in parts of Africa as evidence that governments are becoming more assertive, compelling multinational tech firms to accommodate local concerns more frequently than they might otherwise anticipate.
Infrastructure Gap and Data Centre Limitations
Leverage in the AI era is not solely political; it is also infrastructural. Africa remains dramatically under‑represented in the global digital economy’s physical backbone, accounting for less than one percent of worldwide data‑centre capacity despite housing roughly 18 percent of the world’s population. Research by McKinsey shows that the combined capacity of the continent’s five largest data‑centre markets falls short of that of France. Persistent electricity shortages further constrain expansion, making negotiations over data‑centre and cloud infrastructure especially sensitive and high‑stakes.
Kenya’s Contested $1 bn Data Centre Deal
One of the most closely watched projects is a proposed $1 billion data‑centre development involving Microsoft and Emirati firm G42 in Kenya. President William Ruto drew attention to the project’s massive energy demands, warning that such a facility would require substantial additional power generation. Reports have also highlighted discussions over commercial terms and long‑term commitments tied to computing capacity. Kenyan officials maintain that talks are ongoing, illustrating the trade‑off governments face: attracting AI infrastructure investment while balancing energy needs, financing costs, and long‑term strategic dependence.
Diversification, Partnerships, and Developmental Impact
The question of who builds Africa’s digital future extends beyond Western tech firms. Sanusha Naidu, a senior research fellow at the Institute for Global Dialogue, observed that debates about diversification—whether shifting from Western to Chinese providers—are ultimately part of a cost‑benefit calculation. She stressed that governments must assess what is returned through these partnerships, noting that data centres are major consumers of water and can exacerbate socioeconomic strains. Drawing parallels with 1990s textile investments, she warned that heavy subsidisation by recipient countries often accompanies foreign inflows.
Data Sovereignty, Surveillance, and Public Trust
Concerns about dependence stretch beyond data centres to the broader suite of foreign‑built digital systems African governments have adopted over the past decade—cloud platforms, e‑government services, surveillance networks, and smart‑city technologies. Concurrently, debates over data governance, digital sovereignty, and where sensitive information should be stored and processed have intensified. Drawing a parallel with calls for an Africa‑led Credit Rating Agency, experts argue that African nations must retain control over their data to protect sovereignty and ensure that technological advances serve public interests rather than external agendas.
Towards Balanced Participation: Local Capacity and Future Outlook
While much of the AI governance dialogue remains confined to policymakers, regulators, and industry actors, Afrobarometer’s Joseph Asunka cautioned that elite‑level negotiations risk alienating citizens. He warned that a trust gap could undermine adoption of fintech, e‑commerce, and e‑government tools. Nonetheless, initiatives such as the UNDP’s timbuktoo programme aim to nurture local innovation, entrepreneurship, and digital infrastructure. Although Africa is unlikely to achieve AI self‑sufficiency, the goal is to shape the terms of engagement so that the continent participates not only as a consumer but also as a contributor to the global AI landscape.