Upskilling Africa’s Workforce for AI: Who Is Responsible?

Every few months, a new number arrives to remind Africa how far behind it sits in the global AI economy. The latest one is stark: fewer than 3 percent of the world’s AI workforce is based on the continent, even as Africa’s AI market is projected to grow past 35 percent annually through 2030. Demand is racing ahead of supply, and the gap is no longer an abstract policy concern. It is showing up in hiring boards from Lagos to Nairobi to Johannesburg, where roles for data scientists, machine learning engineers, and AI product managers sit open for months at a time.

The question that follows is less about whether Africa needs an AI-literate workforce that much is settled and more about who should be paying for it, designing it, and delivering it. Governments, universities, employers, and big technology companies all have a stake in the answer, and each has, so far, done only part of the job.

A Continent Growing Faster Than It Can Train

The scale of the mismatch is worth sitting with. The African Union projects the continent’s youth population will grow by roughly 450 million by 2035, a demographic wave that could become a workforce advantage only if training systems keep pace. Yet the UN Economic Commission for Africa notes that only Mauritius, Egypt, Morocco, and South Africa are currently considered “AI-ready,” leaving most of the continent, Nigeria included, without the infrastructure or institutional readiness to absorb AI at scale.

South Africa’s experience is instructive because it is often held up as one of the continent’s more advanced markets. A 2026 Mercer report found South African employers caught in what it calls a “talent paradox,” where AI adoption is accelerating even as the skilled workers needed to make use of it remain scarce. Academics interviewed by TechCabal have pointed to a structural mismatch behind this: university curricula typically run on five-year review cycles, while the tools graduates are expected to use change every few months. If a country with South Africa’s relatively strong tertiary education system is struggling to keep pace, the challenge for less-resourced systems is considerably steeper.

What Nigeria Has Tried

Nigeria’s answer, at least on paper, is the largest state-led digital skilling effort on the continent. The 3 Million Technical Talent programme, run by the National Information Technology Development Agency, set out in December 2023 to train three million Nigerians in technical skills, including AI and data science, by 2027. NITDA reports that the programme now has a presence in all 774 local government areas and has conducted more than 60,000 psychometrically validated assessments, with a 99.6 percent completion rate. Amazon Web Services, one of the programme’s infrastructure partners, has reported that more than 7,500 fellows had secured jobs through employer networks and gig placements tied to the initiative.

Those figures are genuinely significant for a government-run programme in a country where digital infrastructure is uneven. But completion rates and assessment counts measure participation, not employability, and the harder question — how many fellows are in AI-adjacent jobs a year after finishing, and at what salary — is not yet answered with the same clarity. NITDA has since layered on additional tracks, including a DeepTechReady stream for advanced AI and data science, and a five-year partnership with the European Union aimed at deepening technical training beyond entry-level digital literacy.

The Private Sector’s Uneven Contribution

Big technology firms have not stayed on the sidelines, but their involvement tends to follow their own commercial interests rather than a coordinated national plan. Microsoft has committed to training one million South Africans in AI skills and funding 50,000 certification exams, while its Africa Development Centre supports developer ecosystems more broadly. Huawei’s ICT Academy partnership with Ahmadu Bello University has grown from two certified instructors to more than twenty, a model now being replicated at other Nigerian institutions. Within the 3MTT partner network, a “global technology company” is credited with funding AI and data science upskilling for 20,000 Nigerian youth and helping shape curriculum against international standards.

These programmes are useful, but they are also fragmented. Each vendor trains toward its own certification stack, on its own timeline, with its own definition of what “AI-ready” means. A fellow certified under one company’s programme may find that certification carries little weight with an employer using a different toolchain. Without common standards, the private sector’s contribution risks becoming a patchwork of overlapping, sometimes competing, credentials rather than a coherent talent pipeline.

Employers Have Skin in the Game Too

A recurring finding across recent workforce surveys is that employers themselves increasingly accept responsibility for reskilling, even as they hedge on job security. One survey cited by TechBuild Africa found that 77 percent of employers plan to upskill workers in response to AI, while 41 percent still expect to reduce headcount in roles where tasks can be automated. That tension matters for how training is designed. If AI tools increasingly handle the repetitive tasks junior employees once used to learn a trade, then employers who want a skilled mid-career workforce in five years need to be deliberate about how junior staff build judgment, not just tool fluency. A data analyst still needs to recognise unreliable model output. A developer still needs to review, not just accept, AI-generated code.

No Single Actor Can Close This Gap Alone

Set against that backdrop, the honest answer to “who is responsible” is that no single actor can close this gap by itself, and the search for one is arguably part of the problem. Governments can fund scale and reach — 3MTT’s footprint across every Nigerian local government area is proof of what state coordination can achieve — but government programmes struggle to keep training content current at the pace AI tools evolve. Universities carry the credibility and depth needed for foundational skills, but their curriculum cycles are simply too slow to track a field moving in months. Private technology firms move fast and bring real technical expertise, but their programmes are shaped by commercial incentives and rarely designed for interoperability with each other. Employers sit closest to the actual skills gap in their own operations, yet most have shown they will upskill selectively rather than systemically.

What is missing across most African markets, Nigeria included, is not effort but coordination: a shared framework for what AI competency actually means at each career stage, recognised across government, university, and employer credentials alike. Until that exists, the continent will keep producing pockets of well-trained talent that struggle to translate into the broad-based, employable workforce the moment actually requires.