AI Governance and National Strategies: Shaping Africa’s Digital Future
By Estevão Zinga
Almost four years have passed since ChatGPT was launched on 30 November 2022. In that short span, artificial intelligence has left the laboratories of California and Shenzhen and added a new dimension to the economy: the ability to turn knowledge and data into systems that learn, make predictions, generate content and support decisions. This has changed the nature of international competition. Governments have realised that relying entirely on foreign infrastructure, models and services can create economic, technological and institutional vulnerabilities.
UN Trade and Development (UNCTAD) projects that the global AI market will grow from $189 billion in 2023 to $4.8 trillion in 2033. According to the Stanford AI Index 2026, private AI investment in the United States reached $285.9 billion in 2025. And the International Energy Agency projects that electricity consumption by data centres will nearly double, from 485 TWh in 2025 to around 950 TWh in 2030. For Africa, the question is no longer whether AI will shape its economies, but who will shape AI on the continent, and on whose terms.
A December 2025 report by the African Development Bank estimates that inclusive AI adoption could add up to $1 trillion to Africa’s GDP by 2035, nearly one-third of the continent’s current output. It is a scenario, not a promise, and it is precisely this conditionality that makes governance decisive. Africa’s greatest risk is not a lack of vision, but what I call the execution deficit: the gap between what strategies promise and the capacity to deliver them.
Governance is not the same as regulation
Public debate often confuses governance with regulation. Regulation is only one instrument: alongside the law there are principles (OECD, UNESCO), technical standards (ISO/IEC 42001, NIST AI RMF), audit mechanisms and non-binding declarations. Governance is the broader system through which a country decides what it wants from AI and organises itself to achieve it. A country can pass an AI law and still lack governance. What it cannot do is govern without capability: governments cannot effectively govern technologies they do not understand, monitor or have the capacity to audit.
Governments hold a distinctive place in this system: it falls to them, first and foremost, to align ministries, set long-term priorities and protect citizens’ rights. But they do not act alone. Companies invest and innovate, universities produce knowledge and talent, and civil society holds institutions to account. The state’s task is to lead and coordinate this ecosystem, not to replace it.
Lessons from the major AI powers
The United States, China, the European Union and India pursue different strategies, but with common priorities: computing power, energy, data, talent, innovation and applications.
The European Union chose rules. The AI Act, in force since 2024, assigns obligations according to risk, but implementation has proved demanding: in 2026 the “Digital Omnibus” postponed obligations for high-risk systems to December 2027. Ambitious regulation requires technical standards and institutional capacity.
The United States went the opposite way. “America’s AI Action Plan”, published in July 2025, bets on speed, deregulation and an unrivalled volume of private capital.
China combines state direction with industrial scale: its “AI Plus” initiative, issued in August 2025, targets a penetration rate above 70% for intelligent terminals and AI agents in key sectors by 2027, and above 90% by 2030.
India may be the most relevant reference for Africa. The IndiaAI Mission, approved in 2024 with around $1.25 billion over five years, treats computing power as shared infrastructure: more than 38,000 GPUs are available at subsidised rates to start-ups, researchers and public bodies. Yet the Indian case also shows the execution challenge: by April 2026, only about $48 million of the $1.25 billion approved had actually been released.
None of these models can be transplanted, but all of them show that the leading AI powers have matched their vision with institutions, funding and measurable targets.
Where Africa stands
Africa has made progress in setting direction. The African Union’s Continental AI Strategy, adopted in 2024, provides a common framework, and the Africa Declaration on AI was signed in Kigali in 2025. According to African Union officials, 17 countries, about one-third of the continent, had adopted national AI policies by September 2026, although counts vary by source. In two years, the number of countries with AI policies has more than doubled, from 8 to 17. The relevant question is no longer how many countries have a strategy, but how many can fund and execute it.
This is where the execution deficit becomes visible. A national AI strategy without funding, institutional ownership and measurable outcomes is a declaration of ambition, not a development plan. According to the International Telecommunication Union (ITU), only 36% of Africa’s population used the internet in 2025, and the region has the widest urban-rural gap: 55% of urban residents are online, compared with 21% in rural areas. The African Development Bank notes that Africa accounts for less than 1% of global data-centre capacity and that around 3 million professionals would need to be trained for AI-related jobs. With just over one-third of its population online and less than 1% of the world’s data-centre capacity, Africa risks consuming AI designed elsewhere, trained on data that is not its own. This skills gap is a reminder that Africa’s greatest AI asset will not be its algorithms alone, but the people capable of understanding, developing and applying them.
From legislation to strategy: Angola and Rwanda
Many African countries face the same dilemma: legislate first, or build capacity first. Angola chose the legislative route. Its ICT White Paper 2023–2027 provides for a National Artificial Intelligence Strategy. In 2025, the ministry responsible for ICT put a draft Artificial Intelligence Law to public consultation, and the bill has since moved to the National Assembly, where it has yet to be approved. The country has also launched UNESCO’s AI Readiness Assessment and authorised the purchase of an AI platform for public administration, financed by the World Bank.
Legislation, however, is no substitute for strategy. When the law arrives before capacity, the state creates obligations it cannot enforce, and a rule that is not enforced loses authority. A law defines what is permitted; a strategy defines what a country intends to achieve, who is responsible and how it will be financed. Rwanda followed the opposite sequence: it adopted a data protection law in 2021, a national AI policy in 2023 and, in June 2026, approved the creation of a National AI Agency, without yet having a general AI law. Even there, the execution deficit persists: analysts note that the policy has no dedicated budget line. For Angola, and for the many African countries now preparing AI legislation, the priority should be to complete the national strategy and align it with the legal framework, so that regulation does not advance faster than the capacity to implement it.
Four strategic pillars
National AI capacity rests on four mutually reinforcing pillars:
- Infrastructure and compute. Reliable energy, affordable connectivity and access to computing power. The next digital divide may be defined not only by who has internet access, but by who has the computing power to build with AI.
- Interoperable public data, credible data protection and datasets in African languages. If African realities are missing from the data, they risk being missing from the decisions made by AI.
- Talent and research. Digital literacy as the foundation, specialised skills and university-industry partnerships that keep talent on the continent. Nations that only consume AI will depend on the knowledge of others; nations that invest in research can help create the knowledge of tomorrow.
- Institutions and trust. A coordinating authority, proportionate risk-based rules, responsible public procurement and transparency towards citizens and investors.
The pillars depend on one another. Compute without data sits idle; data without skills creates little value; and none of it scales without trusted institutions. Some will object that, while millions of Africans still lack electricity, AI is a luxury. But the choice is not between energy and AI: energy, connectivity and data are the foundation of both, and AI can raise the return on services that states already provide.
An execution agenda for Africa
The effectiveness of a national artificial intelligence strategy should be judged by its ability to turn political priorities into concrete results. To that end, African governments should answer five strategic questions:
- Strategic alignment and national priorities: Is the strategy tied to national development goals and focused on the sectors where AI can create the most value?
- Institutional leadership and governance: Is there a body with the authority and accountability to coordinate delivery across ministries?
- Funding and capacity building: Is there sustainable public and private funding for the four pillars: infrastructure and compute, data, talent and research, and institutions and trust?
- Portfolio and benefits management: Is the strategy organised as a sequenced portfolio of programmes, with measurable targets and verifiable economic and social benefits?
- International cooperation and strategic autonomy: Does cooperation with regional and international partners build national capability and reduce critical technological dependencies?
These five questions make it possible to assess the coherence between political ambition, institutional capacity, available resources and effective delivery. A strategy without clear leadership, sustainable funding, programme coordination and evaluation mechanisms risks remaining a statement of intent.
A national AI strategy only becomes a development strategy when its results start to be felt in schools, hospitals, agriculture, industry and the wider economy.
Conclusion
Africa does not need to win a global AI race to benefit from AI. It needs to decide, country by country, what AI should achieve, and to build the institutions capable of delivering it. Governments must lead that effort, not because the state should do everything, but because no other actor is as well placed to set direction, align stakeholders and protect the public interest at the same time. In 2035, the difference between the African countries that capture the AI dividend and those left behind will not be explained by access to technology, which is becoming widespread, but by the quality of the institutions that govern it. In the end, Africa’s AI future will not be determined by how quickly it adopts technology, but by how effectively it turns that technology into knowledge, productivity and opportunity.
References
- African Development Bank (2025). Africa’s AI Productivity Gain: Pathways to Labour Efficiency, Economic Growth and Inclusive Transformation. December 2025.
- UNCTAD (2025). Technology and Innovation Report 2025: Inclusive Artificial Intelligence for Development. April 2025.
- Stanford HAI (2026). The AI Index Report 2026. April 2026.
- International Energy Agency (2026). Key Questions on Energy and AI. April 2026.
- European Union. Regulation (EU) 2024/1689 (AI Act); Regulation (EU) 2026/1744 (Digital Omnibus on AI), in force since 27 July 2026.
- The White House (2025). Winning the Race: America’s AI Action Plan. 23 July 2025.
- State Council of the People’s Republic of China (2025). Opinions on Deepening the Implementation of the “AI Plus” Initiative. August 2025.
- Government of India, PIB (2025). IndiaAI Mission backgrounder, 30 December 2025.
- African Union (2024). Continental Artificial Intelligence Strategy. July 2024.
- Rwanda Ministry of ICT and Innovation (2023). The National AI Policy.
- Onunwa, G. and Shehu, M. (2025). State of AI Policy in Africa 2025. Column Content, October 2025.
- First EthicalAI (2026). AI Governance Frameworks: The Seven Types That Exist. 29 June 2026.
- ITU (2025). Measuring Digital Development: Facts and Figures 2025.
- Government of Angola, MINTTICS. ICT White Paper 2023–2027 (Livro Branco das TIC); Draft Law on Artificial Intelligence (public consultation, 2025).
- TechAfrica News (2026). The AU’s Question: What 55 Countries Need to Agree on Before Africa Can Govern AI. 29 September 2026.
- MediaNama (2026). IndiaAI Mission: Rs 400 Cr Released of Rs 10,372 Cr Plan. April 2026.
- iAfrica (2026). Rwanda Approves National AI Agency. June 2026.
- The Democratic Republic of Congo, the world’s top cobalt producer is lifting its export ban starting October 16th. Congo will now control global supply through strict annual export quotas, the country’s strategic minerals regulator announced Sunday.
- Visa Denied, History Delayed: Somali Referee’s World Cup Dream Sparks Global Debate
- Fifteen Voices, One Legacy: How Lagos Fashion Week Shaped Africa’s Creative Vanguard
- Economic Community of West African States (ECOWAS) Aviation Ministers Set to Address the High Cost of Air Transport and Travel Tickets through Common Reforms in Lomé, Togo
- Ugandan Troops and Somali National Army Recapture Strategic Villages from Al-Shabaab.
- Africa: Tanzania marks record agricultural achievement as African Development Bank President Adesina urges investment in Africa

