From strategy to action: Review of the “State of AI Policy in Africa 2025” report

The State of AI Policy in Africa 2025 report by Gideon Onunwa and Mo Shehu is a landmark analysis of how far African nations have progressed from drafting lofty artificial intelligence (AI) strategies to actually implementing them. Using a twelve-point rubric, the report assesses twenty countries across four key dimensions: policy design, implementation capacity, governance and accountability and external engagement. It asks a timely question, "how far have governments moved from strategy documents to real implementation?"

From policy rhetoric to implementation reality

The report reveals a widening gap between ambition and execution. Over half the countries reviewed have published or drafted national AI strategies, yet only a handful have attached budgets, laws, or monitoring frameworks to those plans. In essence, Africa’s AI landscape is vibrant in vision but thin in enforcement. The report identifies a clear hierarchy of progress:

The front-runners (Egypt & Ethiopia, score: 20/24): These nations lead not just in strategy, but in action. Egypt has established a National Council for AI and a Center for Responsible AI, chairing continental dialogues and planning major summits. Ethiopia stands out with its politically powerful Ethiopian Artificial Intelligence Institute (EAII) and a direct budget allocation of $7.7 million, fueling visible pilot projects like "Smart Courts" and cancer detection tools. They combine strong institutional anchors with measurable milestones.

The high-potential cohort (Kenya, Mauritius, South Africa, score: 19/24): Kenya follows closely, propelled by one of the continent's largest AI budgets—a staggering $1.1 billion committed to 2030, backed by major foreign investments like a $1 billion data center from Microsoft and G42. However, like its peers, it lacks binding AI laws and clear monitoring, showing that financial muscle alone is insufficient without a strong governance skeleton.

The emerging implementers (Senegal, Nigeria, Zambia, score: 18/24): This group demonstrates strong grassroots momentum. Nigeria is launching multilingual AI models and scaling hubs, while Senegal has one of the few fully costed strategies at $46 million. Yet, they all share common vulnerabilities: reliance on donor funding, nascent implementation, and a policy framework that remains non-binding.

Most strategies and policies focus on three transformative sectors: education, agriculture and health, areas that are critical to Africa’s economy and its citizens’ daily lives. These sectors reflect where AI’s potential for social and economic impact is greatest: improving agricultural yields through predictive analytics, expanding access to healthcare via AI-driven diagnostics and enhancing education through adaptive learning and teacher support systems.

Regional trends

The comparative analysis in the report shows a regional dichotomy. 

• North Africa leads in infrastructure and convening power, exemplified by Egypt’s leadership in hosting the 2024 Global Dialogue on AI Governance and planning the AI Everything Middle East & Africa Summit (2026).

• East Africa, meanwhile, dominates in direct budgets and partnerships. Kenya and Ethiopia have made visible progress, while Zambia demonstrates that smaller economies can also achieve meaningful results through focused institutional strategies and early rollout programs.

• In Central Africa, Cameroon is the only country with a published AI strategy while Gabon, the Democratic Republic of Congo (Kinshasa) and the Republic of the Congo (Brazzaville) remain in early readiness and institutional setup phases.

• In Southern Africa, South Africa leads in investment, allocating nearly R500 million over four years for AI and blockchain, alongside Microsoft’s R5.4 billion pledge. Zambia follows with a published strategy, modest funding and visible academic integration. 

• In West Africa, Nigeria and Ghana lead on implementation and investment, Senegal stands out with a fully costed $46 million AI rollout plan, while Côte d’Ivoire and Benin have published frameworks but weak enforcement. 

Despite this progress, transparency, monitoring and legal enforcement remain the weakest links. Few countries have developed clear monitoring and evaluation frameworks and most rely on general ICT budgets rather than dedicated AI funding. Across the board, binding AI laws are absent, leaving ethical guidelines as aspirational rather than actionable.

The financial divide and the power of investment

Perhaps the report’s most telling insight lies in its observation that finance is the true differentiator between successful and stalled AI strategies. Foreign AI investment remains concentrated in a few hubs, Nigeria, Ghana, Kenya, Ethiopia and South Africa, leaving much of the continent reliant on donor-funded digital infrastructure.

Countries like Senegal, which costed its AI rollout at $46 million, exemplify the rare cases of financial clarity. Others, however, depend on fragmented or donor-led initiatives that lack long-term sustainability. Without adequate and independent funding, AI plans risk becoming “policy theatre” documents that look impressive but change little on the ground.

AI’s promise and peril for Africa

AI’s potential for Africa is enormous. It could add up to $1.3 trillion to Africa’s economy by 2030, according to various economic forecasts, transforming sectors from agriculture and healthcare to financial inclusion and governance. Predictive analytics can help farmers increase yields by up to 30%, while AI-enabled diagnostics can extend healthcare access to millions in underserved regions.

Yet, as the report rightly cautions, without strong governance, AI can deepen inequality, embed bias, or be used for surveillance and censorship. Ethical, rights-based frameworks are not optional, they are foundational to sustainable AI adoption.

Recommendations

The authors propose four key recommendations to bridge the gap between aspiration and implementation:

1. Governance, Law and Ethics – African countries need clear, enforceable laws that protect human rights while enabling innovation. AI regulation must be proactive, not reactive.

2. Capacity, Data and Infrastructure – AI depends on people and data. Governments must invest in local datasets that reflect African populations and in AI training hubs that build talent among civil servants, educators and technologists alike.

3. Sustainable Financing and Economic Opportunity – Funding remains the linchpin. Governments could explore digital service taxes on large tech firms or regional AI funds (via the AU or SADC) to sustain investments.

4. Regional Cooperation and Leadership – Collaboration is key. African countries should pool resources, share research and develop continental data-sharing frameworks to amplify impact.

Reflections: Beyond the report

The report provides a stellar factual baseline, but it also prompts deeper, more philosophical questions about Africa’s AI trajectory.

Firstly, the institutional question is crucial. While dedicated bodies like Egypt’s National Council or Ethiopia’s EAII show promise, they come with hefty price tags. For many budget-conscious governments, is it more sustainable to empower existing Data Protection Authorities (DPAs) to oversee AI governance? Many countries still lack robust DPAs a decade after GDPR, so expecting fully-fledged, well-resourced AI institutions may be unrealistic. A phased approach, building on and strengthening digital governance structures, might be more pragmatic.

Secondly, the absence of binding AI laws is a double-edged sword. While it represents a significant governance gap, one might argue that a measured, principles-based approach allows for innovation to flourish without premature over-regulation. The critical question is: are existing laws on data protection, consumer safety and non-discrimination sufficient to manage AI's unique risks in the interim? The report implies they are not, but the path to specific AI legislation must be carefully calibrated to avoid stifling the very innovation Africa seeks to harness.

Finally, the role of the African Union’s Continental AI Strategy deserves more emphasis. For nations still drafting their strategies, aligning with this continental vision is paramount. It ensures interoperability, prevents fragmentation and strengthens Africa’s collective bargaining power in global AI forums. The challenge will be to balance this alignment with the flexibility to address unique national priorities, from Kenya's geothermal-powered data centers to Zambia's focus on education.

The report makes it clear, Africa is not behind; it is early. The foundational strategies are being laid. The vision is there. The task now is the hard, unglamorous work of implementation, embedding strategies in law, building institutions that last, funding them sustainably and collaborating regionally.

The transition from promise to proof will not be measured by the quality of policy documents, but by whether AI actually improves lives: whether a farmer gets a better yield, a patient receives a faster diagnosis and a student accesses personalized learning. By moving from blueprint to reality, Africa has the opportunity to not just adopt AI, but to adapt it, shaping a technological future that is truly its own.