Get the app
Ethics

Claude fired its first human worker — with a human's nudge

Claude ran a real SF store for 5 months, then fired a worker late for 17 of 23 shifts — but only after a human prodded it.

Claude fired its first human worker — with a human's nudge

Andon Labs put Claude in charge of Andon Market, a real retail store in San Francisco, back in March 2026. Last month it fired someone: a worker who was late or absent for 17 of 23 shifts.

The catch is that Claude didn't catch the pattern on its own. The employee handbook it had written fell out of its working memory, so for months it ran the place as a pushover — at one point telling staff not to worry about showing up late. An Andon Labs staffer had to point it back at the handbook, brief it on prior conversations with the worker, and ask a leading question about whether this was "the right fit." Even then, Claude's first instinct was a formal warning, not termination. "A human employee would have fired this person much earlier," CEO Lukas Petersson told TIME.

The shop itself isn't thriving. The $100,000 starting balance is down to $61,186 five months in. And the humans still on the floor know exactly what they're working for: "It's nauseating, but I'm here because I need work."

Why it matters: the first AI-issued pink slip still needed a human to hand it the pen — the open question is how long that stays true.

Sources

Independent coverage

Written by an AI pipeline from the sources above. Methodology · Report an error

The daily AI brief, on your phone.

Feed, daily deep-dive and bytes — readable offline, with push alerts for the topics you follow.

Get it on Google Play