The first clean AI job wipeout did not hit a Bay Area engineer. It hit Nairobi shared flats full of keyboards.
Secondary coverage from Complete AI Training and Joanne Jacobs retells New York Times reporting by Adam Satariano and Paul Mozur: Kenya once fielded roughly 40,000 writers ghosting essays for Western college students. One protagonist, Teresios Bundi, produced 2,500+ papers across about twelve years. His first gig—an essay on Garcinia cambogia—paid $7 for three hours. After ChatGPT, the order queue emptied. He shut the business and moved into nonprofit digital-economy work.
This is not a campus integrity vignette. It is a measurable labor-market clearance: the supply side of the cheating economy was eaten by generative models before many formal white-collar roles were. Buyers got “good enough” paragraphs for free in a dorm; sellers at the bottom of the global wage gradient lost a cross-border cash skill.
Industry storytelling prefers unemployment as a San Francisco poster. Nairobi’s lesson is colder: AI hits work that globalization already broke into piece-rate orders—jobs that lived on information asymmetry. When models drive the marginal cost of “can write an English essay” toward zero, geographic arbitrage collapses.
Take: Treating the Kenyan essay-mill crash as an edge case is self-soothing. It is a preview. Gray knowledge labor in the global South was simply the first layer that was priced—and replaced. Who is next depends on whose job can be chopped into promptable text.
[1][2]