More than 80% of African countries are so far behind in artificial intelligence (AI) preparedness that they don’t even rank in global readiness categories, according to the Global AI Index. Only Egypt, Nigeria, and Kenya are considered “nascent,” while Morocco, South Africa, and Tunisia are “waking up.”
By contrast, the United States and China remain the only “power players,” dominating global AI development, funding, and influence.
Several African nations – including Rwanda, Ghana, and Senegal – have drafted AI strategies, but experts warn that many are overly dependent on foreign models. At the recent NADPA conference, experts stressed the importance of building local, context-specific policies to ensure AI serves African development needs rather than importing global biases.
“Africa must not outsource its AI future,” said Ikenna Ikeme of MTN Nigeria. “Using foreign-made models disconnected from our languages and values could deepen inequality.”
The copy-paste approach to policy is not new. Nigeria’s 2019 Data Protection Regulation (NDPR) closely mirrors the EU’s GDPR, helping kick-start local awareness around data privacy. But analysts argue that the NDPR lacks strong enforcement and overlooks local priorities, like children’s rights and automated decision-making safeguards.
Mozilla tech policy fellow Kiito Shilongo cautioned that the biggest risk is not borrowing from other models, but developing frameworks without input from the communities they aim to protect.
As of mid-2025, only nine African countries have formal AI strategies. Another nine are in the process. Mauritius led the way in 2018, while Kenya formed a national task force on AI and blockchain the same year. Still, many countries remain unranked, under-resourced, or dependent on imported models that may not reflect African realities.
Experts call for a more localized approach – one that balances global standards with African-specific challenges, including informal economies, data scarcity, and language inclusion. Without it, AI could widen existing digital divides rather than bridge them.