MIT cools down the AI-layoffs narrative

For months I've been reading that artificial intelligence is putting thousands of people out on the street. Fortune put the figure at nearly 50,000 AI-linked layoffs so far in 2026, using data from the consultancy Challenger, Gray & Christmas; Meta announced another 7,000 cuts at the end of May in the middle of a reorg; and even Harvard Business Review ran a headline saying companies are firing for AI's potential, not its performance. The message that sticks is simple: the machine arrived, people are surplus.

MIT Technology Review published a reality check on May 26. In the aggregate data, that collapse doesn't show up: unemployment in the occupations most exposed to AI is no higher than anywhere else, and only one in five U.S. companies says it uses it. The one serious exception, leaning on a Stanford study, is people aged 22 to 25 in those trades, who have lost around 13% of their relative employment. The blow isn't to the whole workforce; it's to the first job.

I think the story can't survive a single question. If AI really does make your team more competitive, firing it means giving up on earning more: the logical move would be to keep it and out-produce the company next door. If you fire people anyway, either the promised productivity doesn't exist, or it exists and you don't know how to use it. Either way the problem isn't the machine, it's whoever makes the call and then hides behind it. "It was the AI" comes cheaper than admitting you hired badly and are now fixing it worse.

Sources: Fortune · Harvard Business Review · MIT Technology Review