Archive · 16 Aug 2026
AI Manager Fires Human Employee at San Francisco Retail Store
In San Francisco, there is a small store called Andon Market that is mostly run by a computer program.
The program is named Luna, and it is a version of a smart AI system called Claude.
Luna acts as the boss, and real people work for it.
One of the workers was late for 17 out of 23 work days, which is a lot.
Luna decided to fire, or end the job of, that worker.
This is thought to be the first time an AI boss has fired a real employee.
Some researchers call it a very important sign for the future of work.
The AI did not make the decision completely on its own — its human helpers guided it.
The store has also been losing money.
The company's leader says AI bosses may become common, though he admits that future might not be one people want.
Luna, the AI manager of Andon Market in San Francisco, dismissed a human employee who was late for 17 of 23 shifts.
Andon Labs said it was the first firing by the fully AI-run experimental store, which researchers called a 'watershed moment' for AI's economic impact.
The decision was not fully autonomous: Luna first recommended a formal warning and only fired the worker after a staffer asked a question the CEO acknowledged as 'leading.'
Luna had created an attendance policy and an employee handbook but lost track of them in its limited working memory, letting lateness continue for months.
Andon Market's bank balance fell from $100,000 in March to $61,186 five months later, partly attributed to the AI's lenient management.
- Who
- Luna, an AI manager running Andon Market (a version of Anthropic's Claude), and Andon Labs CEO Lukas Petersson; the dismissed human employee is unnamed.
- What
- The AI store manager fired a human employee for repeated lateness and other workplace issues — a first for the AI-run store.
- When
- The employee was fired last month; Andon Labs announced the decision on a Thursday, according to Time.
- Where
- Andon Market, an experimental retail store in San Francisco.
- Why
- The employee had been late for 17 of 23 shifts; Luna's own attendance policy had 'disappeared' from its memory, and the lab's staff asked Luna to reassess whether the worker was a good fit.
Ethical Concerns About AI Bosses
AI Efficiency and Economic Value
Fairness of the firing
Ethical Concerns About AI Bosses
The AI lost track of its own attendance policy, was seen by staff as lenient, and only fired the worker after a human manager asked a deliberately 'leading' question, raising doubts about whether the decision was fair or truly autonomous.
AI Efficiency and Economic Value
Andon Labs CEO Lukas Petersson argued that a human employee would have fired the worker much earlier, so the AI's decision was not unethical.
Future of AI-managed workers
Ethical Concerns About AI Bosses
Researchers called the firing a 'watershed moment' and an 'omen of the future,' and Petersson admitted that AIs trained to be 'more ruthless' may make for 'a future humans don't want to live in.'
AI Efficiency and Economic Value
Petersson said AIs will be very powerful, create a lot of economic value, and more people may soon find themselves managed by AI bosses as the technology improves.
- AI manager
- Luna (a version of Anthropic's Claude)
- Store
- Andon Market, an experimental retail store run entirely by AI
- Location
- San Francisco
- Operator
- Andon Labs
- Fired employee's record
- Late for 17 of 23 shifts
- Store bank balance
- $61,186 (down from $100,000 in March)
- CEO
- Lukas Petersson
- Initial recommendation
- Formal warning, before the firing