An artificial intelligence agent has recommended ending a human employee’s contract at Andon Market, a San Francisco store operated as an experiment by Andon Labs.
The agent, named Luna, is designed to manage much of the store’s operations, including product selection, hiring, financial oversight and employee evaluations. According to the company, the employee was late for 17 of 23 shifts and continued to show performance issues despite repeated warnings and additional training.
Andon Labs asked Luna to review the employee’s attendance records and assess the case against the store’s policies. The agent initially suggested a formal warning before recommending termination after receiving further guidance from the company, according to the announcement.
Human staff reviewed the recommendation and carried out the decision after checking its legal and ethical implications. Andon Labs co-founder Lucas Peterson said the team would have intervened if Luna had made an incorrect or unethical decision, while suggesting that a human supervisor might have acted sooner.
Luna runs on models from Anthropic and controls an operating budget of $100K. The experiment is intended to test how far AI agents can manage businesses and make decisions affecting human employees, while workers remain employees of Andon Labs with their legal rights and benefits.
Source: Jawlah
