More than half of African governments have published national artificial intelligence (AI) strategies, but only a fraction have attached budgets, laws, or monitoring frameworks to make them work, according to a new study by tech researchers Mo Shehu and Gideon Onunwa.
The report, titled The State of AI Policy in Africa 2025, exposes a widening gap between ambitious policy documents and actual implementation across the continent. While AI is reshaping global economies, Africa’s biggest challenge isn’t innovation but turning written plans into functioning systems that can regulate and harness the technology.
The researchers examined AI policy across 20 African countries using a twelve-point assessment covering institutional setup, funding, legal frameworks, ethics, and monitoring. Their findings reveal uneven progress and what they describe as a clear divide between ambition and execution. Many strategies remain invisible to the public, with limited consultation or published updates that reduce transparency and accountability.
Egypt and Ethiopia lead the continent, scoring highest for strong institutional anchors and visible pilot projects. Both governments have tied AI policy directly to national development goals, an approach the authors characterize as linking technology to continuity. Kenya follows closely with a $1.1 billion AI budget driven by major foreign investment and private sector collaboration, though the report cautions that Kenya and Senegal stand out for scale and financing but lack strong legal enforcement and public monitoring.
That absence of oversight represents a recurring weakness across African AI policies, making them vulnerable to inertia and reducing their real-world impact. The report notes that only a few governments have dedicated, multi-year AI budgets, with most depending on donor support or general information and communications technology (ICT) allocations that are easily diverted. This financial fragility undermines long-term sustainability of AI initiatives.
Ghana, Mauritius, and South Africa show promising momentum, maintaining active consultation with researchers and tech entrepreneurs. However, they remain weak on legal codification and monitoring and evaluation (M&E) systems, limiting their ability to track whether policies are actually working.
Nigeria, Africa’s largest economy, receives praise for grassroots innovation including development of local language AI models. Yet the report faults the country for inadequate financing, noting that Nigeria demonstrates strong participation and implementation through local language models and scaling hubs but lacks dedicated funding and statutory backing.
Zambia emerges as an unexpected bright spot despite its smaller economy. The authors highlight how Zambia shows that smaller economies can still achieve meaningful progress through institutional focus and early rollout, suggesting that political will and strategic planning can compensate for limited resources.
Countries like Namibia, Botswana, and Zimbabwe remain stuck at the conceptual stage, where discussions are largely fragmented across ministries. These nations are conducting readiness assessments or drafting strategy outlines but have yet to publish full national frameworks. For roughly 34 nations including Chad, Eswatini, Sierra Leone, and the Central African Republic, there is little to no government-led initiative or official consultation on AI.
The absence of clear accountability systems represents an equally concerning problem. Most strategies lack indicators or reporting systems to track progress, leaving citizens and stakeholders unable to determine whether national AI goals are being met. While 35 African countries have data protection laws, few have AI-specific regulations or technical regulators capable of auditing algorithms.
Ethics has become what the researchers call a hollow promise. Many national strategies include principles on fairness and transparency, but few back them up with tangible enforcement mechanisms. Governments often include ethical language in policy documents but rarely translate it into binding rules or institutional practice.
Shehu and Onunwa describe the result as ambitious frameworks without the institutional or financial backbone to sustain them. This creates a cycle of ambition without accountability that could have serious consequences for the continent’s technological future.
Without robust oversight or local regulation, African AI ecosystems risk being dominated by foreign corporations and imported technologies, which could entrench dependency rather than drive innovation. Citizens remain vulnerable to algorithmic bias and data misuse in the absence of clear legal protections.
The report argues that AI represents not just a technological revolution but a geopolitical one. Countries that shape the technology also shape the rules, making Africa’s participation essential to ensure its needs and perspectives are represented in emerging global norms. However, opacity rather than lack of talent could stall progress, with transparency, monitoring, and legal enforcement remaining the weakest links.
Despite these challenges, the authors remain cautiously optimistic. Kenya’s billion-dollar AI budget and Senegal’s costed rollout plan suggest political will is beginning to translate into fiscal commitment. These examples demonstrate that meaningful progress is possible when governments move beyond drafting documents to actual implementation.
The researchers propose several solutions: embedding AI strategies in law, strengthening public institutions, making monitoring frameworks public to allow independent verification, and creating regional funding mechanisms to sustain national AI investments. They argue that countries treating AI as a public good rather than a private experiment will shape the continent’s technological and economic future.
Shehu and Onunwa emphasize that Africa is not falling behind but is early in its AI journey. The challenge is making ambition durable by shifting focus from drafting new strategies to implementation, evaluation, and collaboration. The next phase requires governments to prove their policies work rather than simply announcing them.
The risk is that Africa becomes a testing ground for foreign AI systems rather than a co-author of their regulation. The opportunity, however, is immense: building AI governance from the ground up, informed by ethics, inclusion, and development priorities unique to the continent. Whether African governments can translate policy documents into functioning systems will determine whether the continent leads in shaping AI’s future or remains dependent on technologies and regulations developed elsewhere.
