Africa on the IMF's AI Preparedness Index: What the Numbers Show, and What They Miss.
The IMF gave 174 economies a score for how ready they are to use AI. Africa sits at the bottom. The more interesting question though, is whether the test is measuring the right things.
In June 2024, the International Monetary Fund did something useful and slightly dangerous: it put a number on how ready each country is for artificial intelligence.
The AI Preparedness Index, or AIPI, scores 174 economies from 0 to 1 across four pillars ; digital infrastructure, human capital and labour-market policies, innovation and economic integration, and regulation and ethics. It's a serious piece of work, and it has become a reference point in the global conversation about who will benefit from AI and who will be left behind. For anyone thinking about AI in African finance, it's worth understanding what it says. It's worth understanding its limits even more.
Let me do both, in that order. The data first, because it's real and it matters. Then the harder question is whether an index built to measure readiness for a particular kind of AI can actually see what's happening on this continent.
What the numbers show
The headline is a gradient, and it runs exactly the way you'd fear. Advanced economies average around 0.68 on the composite index. Emerging markets sit near 0.46. Low-income countries average roughly 0.32. Sub-Saharan Africa, as a region, lands around 0.34 -last among the world's regions.
At the top of the global table sit Singapore, Denmark, and the United States, with scores near 0.80. Africa's numbers live in a different band entirely. And within the continent, there's a clear internal hierarchy. South Africa ranks highest, at roughly 0.5, followed by a cluster of relative front-runners Kenya, Ghana, Rwanda, Botswana, and Namibia. Notably, Nigeria , Africa's largest economy and most populous nation ranks among the lower performers in the region, a reminder that economic size and AI readiness are not the same thing.
The pillar breakdown is where the diagnosis gets specific, and it's consistent with everything we know about the continent's constraints. Digital infrastructure is the binding one: only about 37% of Africans were online as of 2023, by the ITU's estimate, with something like 900 million people still offline. You cannot deploy data-hungry AI at scale on top of a connectivity base that thin. Human capital tells a similar story Africa graduates a small share of students in STEM fields, and the skills pipeline for building and maintaining AI systems is nowhere near the scale the moment demands.
And then there's the fourth pillar, regulation and ethics which turns out to be the world's weakest, not just Africa's. Roughly 48% of countries globally scored zero on having a national AI policy at all. As the IMF's managing director put it, the area where the world is most lacking is in regulation and ethics. This is the one pillar where the gap between advanced and developing economies is smallest, because almost everyone is failing it. Hold that thought it matters for where Africa's opportunity actually lies.
Taken at face value, the index delivers a sobering verdict. On the infrastructure and skills that a Global North AI rollout would require, most of Africa is, by these measures, underprepared. That's true, and pretending otherwise helps no one. The connectivity gap is real. The skills gap is real. The compute gap is real.
But "by these measures" is carrying an enormous amount of weight in that sentence. And that's where the second half of this piece begins.
What the numbers miss
An index is a theory of what matters, wearing the costume of an objective fact. The AIPI's four pillars encode a specific belief about what AI readiness looks like: dense digital infrastructure, a large pool of STEM graduates, high R&D spending, formal regulatory frameworks. That's a reasonable description of how AI got built in San Francisco and Singapore. It is not obviously the only path, and treating it as the universal yardstick produces at least three distortions when you point it at Africa.
First, it measures readiness for the wrong AI. The index rewards the infrastructure that Northern AI assumes big data centres, ubiquitous broadband, clean and abundant structured data. But the most consequential financial AI on this continent has never depended on those things. It has been built on mobile-money rails that Africa invented and the North largely lacks. M-Pesa turned Kenya into a country where financial data flows through phones rather than bank branches; the AI that matters here reads that data transaction histories, repayment rhythms, informal cash flows not audited filings and credit-bureau files. An index scoring you on your data-centre density and formal-sector legibility will systematically underrate a financial-AI ecosystem built on exactly the opposite foundations. It's grading a fish on its ability to climb!
Second, the scores hide as much as they reveal, and the researchers who've looked closely say so. A 2024 academic study of AI policy across African states found that the global indices misread the continent in both directions. Mauritius scored well on data and infrastructure, yet a close look revealed real weaknesses in its data-protection regime that the pillar score simply didn't capture. Kenya scored lower on governance and yet, the same researchers found, its government has spent considerable political and economic capital building an enabling environment for AI, with genuine strategy behind the number. In other words, a country can score poorly on the regulation pillar while doing more thoughtful regulatory work than a higher-scoring neighbour. The number and the reality had come apart. If you were allocating attention based on the score alone, you'd back the wrong country.
Readers of this publication will recognise the Kenya example, because we've covered it directly: the Capital Markets Authority's 2025 move to bring robo-advisers inside its licensing regime is precisely the kind of concrete, sector-specific AI governance that a high-level "national AI policy" checkbox doesn't detect. Kenya is regulating AI in finance in a real, enforceable way. The index can't see it, because it's looking for a document with "National AI Strategy" on the cover, not a clause buried in capital-markets regulations.
Third , and most importantly the weakest pillar is the cheapest to fix, and that inverts the whole story. The infrastructure gap is brutal precisely because it's expensive and slow: fibre, data centres, and electricity take capital and years. But regulation and ethics, the pillar where the entire world is failing, costs comparatively little to get right. You don't need a national fibre network to write a good data-protection law or a sensible AI-in-finance framework. You need institutional will and clear thinking. This is the one dimension of AI readiness where an African country can genuinely leapfrog not by outspending richer nations on compute, but by out-governing them while the field is wide open. Kenya's financial regulator is quietly demonstrating what that looks like.
How to actually read this index
None of this means the AIPI is wrong or worthless. The connectivity and skills deficits it measures are real, and they are genuine constraints on what's possible. An honest builder doesn't wave the index away; the "digital sand" it exposes is solid ground for a warning. Pretending Africa is secretly AI-ready would be its own kind of dishonesty.
But the index is a diagnosis of readiness for one model of AI the capital-intensive, infrastructure-heavy, data-abundant model of the Global North. Read as a scorecard for that game, Africa is behind, and the index tells you exactly where. Read as the last word on Africa's AI prospects, it misleads, because the AI actually taking root here is playing a different game: lighter on infrastructure, built on mobile-money data, and if its regulators are smart , potentially ahead of the world on exactly the governance pillar everyone else is failing.
So here's how I'd use it. Treat the infrastructure and human-capital scores as a real to-do list; those gaps won't close themselves, and no amount of contextual nuance builds a fiber network. But treat the regulation score as an opportunity flag, not a verdict the pillar where the playing field is most level, the cost of excellence is lowest, and a well-governed African market could lead rather than follow. And treat the composite score with real caution, because it's averaging together a genuine infrastructure deficit with a governance gap that a country like Kenya may already be closing faster than its number admits.
The IMF has told us where Africa stands on the world's chosen measure of AI readiness. The more useful project is deciding which parts of that measure to take as a challenge, which to take as an opportunity, and which to quietly ignore because they're scoring the wrong game. The frontier of AI in African finance won't be built by climbing someone else's index. It'll be built by knowing which rungs are worth climbing.
Frontier Finance AI covers the collision of artificial intelligence and capital markets across Africa and Asia. Index figures are drawn from the IMF's AI Preparedness Index (2024); readers can explore the full dataset on the IMF's DataMapper.
