Two years ago the reasonable criticism of African AI governance was that there was not any. The African Union Executive Council endorsed the Continental AI Strategy in Accra on 18–19 July 2024, and the count of member states with a national strategy could be listed on one hand.
That criticism has expired. By the OECD’s April 2026 case study on AI governance in Africa — drawn from a workshop in Nairobi convening senior officials from twelve countries — the roll of adopters now runs to Mauritius, Tunisia, Ghana, Benin, Senegal, Zimbabwe, Algeria, Ethiopia, Mauritania, Zambia, Côte d’Ivoire, Kenya, Lesotho, Libya and Nigeria, with South Africa, Togo and Uganda in development. Kenya adopted its National AI Strategy 2025–2030 in May 2025. Ghana launched a National AI Strategy running to 2035 in April 2026, built on eight pillars and carrying a stated ambition of GH₵500 billion added to GDP.
This is not a continent that failed to write the documents. It is a continent that wrote them quickly, in a peer environment, and largely to a common template drawn from Agenda 2063 and the UNESCO ethics recommendation.
Then look at the money
Kenya costed the implementation of its strategy at KSh 152 billion across five years. A review by CIPESA of Kenya’s 2026/27 national budget found no dedicated allocation for the strategy at all. The KSh 8.6 billion that went to the ICT sector was directed at broadband expansion, digital skills and the digitisation of government services — all defensible, none of them the strategy.
This is the pattern worth naming. The instrument that a government publishes and the instrument that a government funds are two different objects, and only one of them changes anything.
A strategy with no budget line is not a policy. It is a press release with footnotes.
Why it happens, and why it is not simply cynicism
The temptation is to read the funding gap as insincerity. It is usually something duller and more fixable. AI strategies are written by ministries of ICT, digital transition or technology. Budgets are written by ministries of finance. The two documents are produced on different cycles, by different people, against different
incentives, and the strategy almost never arrives with a costed, sequenced first-year plan that a finance ministry can actually price.
There is a second structural reason. Much of the drafting capacity behind these strategies came from outside — technical assistance from development partners, foundations and consultancies. Zambia’s 2024 strategy, for instance, was developed with technical assistance from the Tony Blair Institute alongside contributions from Finland and USAID. That support is genuinely useful and it accelerates publication. It does not, on its own, create a domestic budget claim, because the officials who will have to defend the line item in a finance committee were often not the officials who wrote the document.
What operationalisation would actually look like
The Charter’s position is that the next phase of African AI governance is measured on four things, none of which is the existence of a strategy document.
First, a named budget line, in the national budget, attributable to the strategy. Not ICT sector spending that could be counted toward it after the fact.
Second, an institution with a mandate and staff. Kenya’s proposed Office of the Artificial Intelligence Commissioner would qualify if it is created and resourced. An inter-ministerial committee that meets quarterly would not.
Third, a working redress mechanism wherever an AI system touches the public. Ghana’s experience is instructive: the Publican AI customs system was deployed at Tema Port in February 2026, and by April a coalition of freight forwarding and trade bodies had struck, citing the complete absence of a functional appeals process. The technology worked. The governance around it did not exist.
Fourth, published implementation reporting. If nobody has to report against the strategy annually, the strategy has no enforcement mechanism, including against the government that wrote it.
The Charter’s read
The adoption race is effectively over and Africa did better in it than the commentary allowed for. The operationalisation race has barely started, and it is far harder, because it requires finance ministries, regulators, procurement officers and courts to do things — not just communications ministries to publish things.
Watch the budget documents, not the strategy launches. For the next twelve months that is where the story is.
Sources line for the foot of the article: OECD, AI Governance in Africa case study, April 2026 · African Union, Continental Artificial Intelligence Strategy, endorsed Accra, 18–19 July 2024 · CIPESA, July 2026 · African Arguments, June 2026.