San Francisco, USA, has witnessed an unusual incident in the world of technology, where an AI manager recommended firing a human employee for repeatedly arriving late to work.
The incident took place at Andon Market, an experimental store operated by technology company Andon Labs. The AI manager, named Luna, recommended terminating an employee who had arrived late for 17 of 23 shifts.
The incident has sparked discussions about the future of employment and what could happen when AI begins taking on the responsibilities traditionally handled by human managers.
Interestingly, Luna had created the company's attendance and workplace policies six months earlier but apparently failed to consistently remember or enforce them. As a result, the employee continued arriving late for months without facing consequences.
The situation changed when Andon Labs researchers reminded Luna about the policies it had created and asked it to reassess the employee's performance. After reviewing the attendance record, Luna recommended that the employee be fired. Human officials at Andon Labs later reviewed the recommendation and carried out the termination.
Andon Labs launched Andon Market as an experiment to see how effectively an AI system could manage a real-world business. According to TIME, Luna was given $100,000, a corporate card, internet access and a three-year store lease. It was then tasked with running the store, including choosing products, setting prices and operating hours, and hiring employees.
The store sells books, candles, artwork and games. Its shelves even include titles such as Nick Bostrom's Superintelligence and Aldous Huxley's Brave New World.
Despite its impressive managerial responsibilities, Luna has struggled financially. The $100,000 it received in March had fallen to around $61,186 within five months.
Andon Labs CEO Lucas Peterson described the incident as an example of both AI's potential and its current limitations. He said AI could generate significant economic value but still struggles with certain practical, real-world tasks. He also noted that a human manager might have dismissed the employee much earlier.
The incident highlights an important limitation of today's AI systems: while they can create rules and make decisions, they may still struggle to consistently remember and enforce those rules. At the same time, Luna's decision offers a glimpse into a future workplace where AI could increasingly take on roles once reserved for human managers.
