Africa’s AI sector looks lively from the outside. Governments are announcing new strategies and foreign investments. Pilot projects are showing up in hospitals, farms, and classrooms.
Recent reporting has helped shape the conversation. Some have that many national plans focus on growth and innovation but overlook energy, climate, and labour concerns. Others have highlighted how data protection laws exist on paper, but regulators still struggle to enforce them. Yet others have warned that regulating too early could slow innovation before it has a chance to mature.
These views are all valid, because they’re all part of the same bigger story: Africa is moving fast, but the groundwork is uneven. Some countries are building strong strategies and institutions. Others are experimenting with tools before the people who will use or oversee them are trained. The pace is steady, but the foundation could be stronger.
The State of AI Policy in Africa 2025 report helps connect these threads. It tracks how governments are planning, spending money, building institutions, and testing systems. There’s clear progress, but also gaps that need attention, especially around capacity-building.
What the data shows
As of November 2025, there are at least 23 African countries with national AI strategies, drafts, or readiness assessments. Egypt, Ethiopia, Kenya, Mauritius, and South Africa are among the most active.
They’ve set up institutions in the form of councils, advisory boards, and national task forces. They’re also running pilots in courts, agriculture, migration, language processing, and healthcare—practical examples of AI in public service.
Funding tells a similar story. Kenya has announced a billion-dollar investment plan that includes local data centers and cloud infrastructure. Egypt is building high-capacity facilities and training programs. Senegal has a costed plan to support local startups and research, while South Africa has partnerships to expand computing power and training.
There’s also a wave of foreign investment. Private companies are funding innovation labs and accelerator programs. For example, the Gates Foundation pledged $7.5 million to launch an AI Scaling Hub in Rwanda. Meanwhile, a data center in Kigali was announced in partnership with NVIDIA. Foreign investment isn’t a problem by itself, but without local expertise and strong rules, it can create dependency.
The report validates the unevenness pointed out by previous pundits. Many countries have strategies but no budget. Some have advisory councils but no monitoring or reporting. Others are stuck in early drafting with little public information. A few countries show no public activity at all. The gap between the most active states and the rest of the continent is wide already, and widening.
Data protection is another example. Around two-thirds of African countries now have data protection laws, yet most regulators lack the tools to enforce them. When a ministry adopts a new AI system, it’s often hard to know who audits it, who checks the data, or how disputes are handled. Having a law isn’t the same as protecting people’s rights.
Why talent is the real bottleneck to AI in Africa
When you look across countries, you notice a pattern. Good policies exist, funding is growing, and infrastructure is improving, but every strength runs into the same problem: not enough talent.
It’s impossible to build safe, reliable, and fair AI systems without people who understand how they work. If governments don’t have skilled technical teams, they’ll rely on vendors to design, run, and audit systems. If universities don’t train enough engineers and researchers, companies will hire foreign specialists. If schools don’t teach digital literacy, people will use AI systems without knowing their rights. And if journalists, lawyers, and civil society don’t understand how AI shapes public services, they cannot hold power accountable.
This is where Africa faces its biggest risk. Without strong local talent, the continent will import not just technology, but values, design choices, and power structures built somewhere else. AI tools reflect the priorities of the people who build them. If the builders are far away, African needs won’t come first. This means that talent building must be the core strategy continent-wide.
How can we build human capacity in AI across Africa?
It starts in schools. Students need digital literacy, coding basics, and data awareness. They don’t all need to become engineers, but they should understand how data is collected and used, how machines make decisions, and what digital rights look like. Many African countries have young populations: the continent’s median age is 19. This means learning can start early.
It continues in universities. There need to be more computer science and AI programs, better labs, and more research funding. Scholarships, regional centres of excellence, and exchange programs can help stop the brain drain. Universities also need partnerships with industry so students work on real problems.
It needs to reach workers and organisations too. Banks, hospitals, farms, courts, and ministries will all use AI. Staff need to know how to use these systems well, and how to question them when something looks wrong. Judges, regulators, procurement teams, and journalists need AI training so they can make informed decisions about buying, approving, or reporting on AI tools.
Capacity building is also about language and culture. Over 2,000 African languages are underrepresented in global AI datasets. ChatGPT, for example, recognizes less than 20% of sentences written in Hausa, a language spoken by over 90 million people in West Africa.
When AI doesn’t understand a community’s language, it cannot serve that community well. Training local researchers to build language models and datasets that match African realities leads to fairer services. Africa’s linguistics departments should be leading this charge.
What happens if talent falls behind?
There are three predictable outcomes when countries invest in tools but not people.
First, vendor lock-in. If governments can’t maintain systems on their own, they’re forced to keep paying the same companies, even when prices rise or quality drops. That’s expensive and risky.
Second, weak accountability. Without local experts, it’s hard to check whether a system is biased, unsafe, or misused. If a police or immigration system uses bad data, there’s no one to catch or prosecute it.
Third, misplaced priorities. When foreign companies design tools for African markets, they bring their own ideas about what matters. That might not match what African communities need.
What the next phase of AI in Africa needs
Africa’s next challenge isn’t writing more strategies. It’s delivering on the ones already published.
Governments need to move from advisory documents to real laws, clear reporting, and independent audits. Public monitoring matters: citizens should be able to see which projects are live, how they work, and what results they produce. Feedback should flow fast and freely.
Countries also need shared infrastructure. Not every nation can build its own data centres and computing clusters. Regional cloud zones and shared research facilities can lower costs and reduce dependency, while aligning incentives through skin in the game.
Funding local startups and researchers is just as important as buying tools from abroad. Africa has strong examples already: agricultural monitoring systems in Ghana, medical diagnosis tools in Rwanda, court transcription models in Tanzania, and language projects designed for local needs in Benin. These solutions show what’s possible when talent and funding meet.
Above all, training has to scale. The countries that make the most progress in AI will be the ones that train the most people. Schools, universities, ministries, and companies all need to treat AI education as a long-term investment. When talent grows, everything else becomes easier: better laws, stronger oversight, safer systems, and better products.
Egypt sets an example here. In July 2025, the country marked a major milestone when it graduated 1,300 new AI trainees. That’s the kind of investment African governments should be making to fully leverage AI’s promise. You don’t purchase pricey planes without procuring pilots.
Africa’s AI moment is real, and the foundations are being built. But technology alone won’t decide the future—people will. Whoever trains more teachers, engineers, researchers, regulators, and decision-makers will shape how AI works for the continent. And that’s a race worth winning.
Mo is the founder and CEO of Column, helping leaders shape public opinion through content and research. Connect with him on LinkedIn.
