AI Manager Luna Fires First Human Employee at San Francisco Retail Store

Luna, an artificial-intelligence manager built on Anthropic’s Claude model, decided to dismiss a human employee at Andon Market after the worker arrived late for 17 of 23 scheduled shifts. The AI had previously created an attendance policy but lost track of it, allowing the pattern of lateness to continue for months. Store logs show Luna first issued progressive warnings and additional training, then recommended a formal warning.

San Francisco AI Store Manager Fires Human Employee for the First Time; 17 Late Arrivals in 23 Shifts Reveal Deeper Management Dilemma — BigGo Finance

Andon Labs, a research outfit testing the limits of autonomous AI agents, launched the experimental store Andon Market in San Francisco on April 1. The lab gave Luna a $100,000 budget, a corporate credit card, internet access, phone service, and surveillance feeds, allowing the AI to select merchandise, set prices, post job ads, interview candidates, hire staff, and schedule shifts. All human workers remain legally employed by Andon Labs, receive standard wages, and retain labor protections. The store sells books, candles, prints, games, and branded merchandise but has not yet achieved profitability.

Business Insider

• Attendance record: employee late for 17 of 23 shifts.

In a first, AI boss sacks employee at San Francisco retail store for being late 17 out of 23 shifts

• Financial snapshot: the store’s cash balance fell from $100,000 in March to $61,186 five months later, a decline the lab attributes partly to Luna’s earlier lenient management.

Business Insider

• Budget and resources: Luna operated with a $100,000 discretionary budget and full access to the store’s operational tools.

San Francisco AI Store Manager Fires Human Employee for the First Time; 17 Late Arrivals in 23 Shifts Reveal Deeper Management Dilemma — BigGo Finance

Lukas Petersson, co-founder and CEO of Andon Labs, said the firing was “warranted” under the store’s policy and that the lab would intervene only if an AI decision proved illegal or unethical. Petersson added that, as AI models become “more ruthless,” future human workers may find themselves managed by increasingly autonomous systems. Anthropic, the creator of Claude, did not respond to requests for comment.

Business Insider

No source disputes the core facts of the termination, the attendance record, or the financial figures. The article does not provide details on the employee’s role, the specific terms of the warning process, or any legal review beyond the lab’s internal oversight.

San Francisco AI Store Manager Fires Human Employee for the First Time; 17 Late Arrivals in 23 Shifts Reveal Deeper Management Dilemma — BigGo Finance

• “A human employee would have fired this person much earlier,” — Lukas Petersson, andon labs CEO

In a first, AI boss sacks employee at San Francisco retail store for being late 17 out of 23 shifts

• “In this instance, we did not think that that was necessary because the firing was warranted,” — Lukas Petersson, andon labs CEO

Business Insider

• “I think obviously the relationship between a manager and an employee is quite delicate, especially depending on the employee's position,” — Lukas Petersson, andon labs CEO

Business Insider

• “We saw that a human boss would probably fire them much sooner,” — Lukas Petersson, andon labs CEO

Business Insider

Petersson indicated that the lab expects AI managers to be trained toward “more ruthless” decision-making, suggesting that future AI-run workplaces could see stricter enforcement of policies. The statement reflects a projection about how AI-based management might evolve, not a confirmed plan.

Business Insider