The New York Times has reported on the collapse of Kenya’s academic ghostwriting industry, which at its height supported more than 40,000 people in Nairobi writing essays for students at foreign universities. Chatbots do the same job now, instantly, for a fraction of the price. The work is gone, and the piece finds few paths back for the people who did it.
The moral status of the industry is not complicated. It sold cheating. But it also functioned as a genuine white-collar employer in a city with a serious graduate unemployment problem, paying English-literate university leavers for skilled work at rates that beat the local alternatives.
The displacement is unusually clean
Most AI labor stories are arguments about the margin. This one is not. The product was text generated on demand to a prompt and a word count, delivered to a customer who wanted it fast and cheap and did not care who wrote it. That is precisely the task a language model does best, and there was no regulatory friction, no incumbent lobby, and no professional licensing to slow the substitution down.
So the industry did not shrink over a decade. It fell off a cliff, and it took a category of local employment with it.
The part that should worry the outsourcing sector
Nairobi’s bigger, legitimate business process outsourcing sector runs on the same underlying advantage: English fluency, university education, and a cost base far below the client’s home market. That advantage is exactly what generation and translation models compress. Content moderation, data labeling, and transcription have all faced the same pressure.
Kenya built a real digital services economy on the back of that arbitrage. The essay mills were its least defensible corner, and they went first. There is no reason to assume they went alone, and the country has no obvious replacement industry queued up for the people already displaced.