Ask someone in Lagos, Nairobi, or Accra what AI means for their job, and you’ll rarely get a neutral answer. It’s either the thing that’s going to replace the call center, the entry-level analyst, the customer service desk, or it’s the thing everyone insists you should be “learning” but nobody explains how. Both reactions come from the same place: nobody wants to be the person left behind.
That fear is reasonable. It’s also, according to the numbers, aimed at the wrong target.
The part that surprised us
The World Bank’s most recent World Development Report put a real figure on this: only about 4.5% of jobs in developing economies, Africa included, face high exposure to automation. In advanced economies, that number is over 14%. The jobs most at risk from AI are, overwhelmingly, not African jobs. They’re the routine, desk-based, already-digitized roles concentrated in wealthier economies.
That doesn’t mean Africa is safe from disruption. It means Africa’s AI story isn’t mainly a replacement story; it’s an infrastructure and skills story. The UN Economic Commission for Africa estimates AI could add up to 3% in economic growth on the continent, but only if the investment goes into the right places: power, connectivity, and, this is the part that involves people like us, skills that are actually built for local problems, not imported wholesale.
Here’s the number that should be driving every conversation about this: Africa’s youth population is projected to grow by 450 million by 2035. That’s not a statistic about the future. That’s a hiring wave that’s already forming. The real question was never “will AI take jobs in Africa.” It’s “who gets to be ready when 450 million young people are looking for work, in an economy AI is already reshaping.”
Ready is not automatic
We’d be lying if we said the continent is set up for this. Right now, only four countries, Kenya, Egypt, Morocco, and South Africa, are considered “AI-ready” by current assessments. There are already more than 2,400 AI-focused organizations working across Africa, which tells you the appetite and the talent are there. What’s missing, consistently, isn’t ambition. It’s the unglamorous stuff: reliable power, affordable bandwidth, and structured pathways for people to actually build the skills before the jobs arrive, not scramble for them after.
What “making jobs” actually looks like up close
This is where it stops being an abstract debate for us at Tech Impact Club Africa, because we’ve watched it happen in rooms we funded.
In Ibadan, a program we’re backing pairs young people with an AI-powered coaching tool for career and business development. Not as a replacement for a mentor, but as something that makes a scarce resource (experienced career guidance) available to people who’d never otherwise get it. In Zambia, a cohort of women who started with zero data background spent eight months learning Python, SQL, and data cleaning, then used it to model real local problems, including food waste, ambulance dispatch, and student dropout risk. None of them lost a job to AI. Several of them are now qualified for jobs that, a year earlier, they didn’t know existed.
That’s the pattern worth paying attention to: the jobs AI is “making” in Africa mostly aren’t AI jobs in the Silicon Valley sense; they’re existing jobs plus a new layer of capability. A career coach who now has an AI tool extending her reach. A public health worker who can now flag risk earlier. A small business owner who can now forecast demand instead of guessing.
So is AI coming for jobs, or making them?
Both things can be true at once, and pretending otherwise doesn’t help anyone. Some routine roles will shrink. But the bigger, faster-moving story in Africa right now is a labor force about to double down on a technology most of the world assumes is only a threat. The data says the opposite: the risk of being automated out is lower here than almost anywhere else, and the risk of being left unprepared for what’s being built is much higher.
The gap between those two outcomes isn’t decided by the technology. It’s decided by who gets trained, and when. That’s the part we can actually do something about and it’s why every program we fund starts with exactly that question: not “how do we adopt AI,” but “who’s going to be ready for it.”