The AI manager handles decisions ranging from staffing and stock selection to scheduling and operations #
An artificial intelligence manager has reportedly made its first firing decision at an experimental shop in San Francisco, dismissing a human employee after they arrived late for 17 of 23 shifts.
The unusual incident took place at Andon Market, a retail experiment operated by Andon Labs in which an AI agent named Luna is responsible for much of the business's day-to-day management. The shop has human employees, but Luna handles decisions ranging from staffing and stock selection to scheduling and operations.
The employee's repeated tardiness eventually became too much of a problem, although the firing was not quite as autonomous and had to have human intervention eventually.
AI Manager Luna Decides To Fire Human Worker #
According to reports about the experiment, the employee had arrived late for 17 of 23 shifts. Luna had previously created an attendance policy, but the AI's memory limitations meant that it did not consistently enforce the rules.
Andon Labs eventually prompted Luna to review its policies and reconsider whether the employee remained a good fit for the shop. The AI initially recommended a formal warning rather than immediate dismissal. It was only after being asked to think more carefully about whether the worker was actually suitable for the role that Luna recommended that the company 'part ways' with the employee.
Human staff subsequently reviewed and carried out the decision, meaning Luna did not independently terminate the person's employment. That distinction has not stopped the incident from attracting attention, however.
The idea of an AI boss effectively telling a human employee that they are too unreliable to keep their job has produced an unusual role reversal at a time when concerns about AI replacing human workers are already widespread.
AI Bosses Could Become More Common #
Andon Market was created as an experiment to see what happens when an AI system is given genuine responsibility over a real-world business.
Andon Labs gave Luna access to money, the internet, business tools and a corporate credit card, allowing it to make decisions that would normally be handled by human managers. Luna even helped recruit the shop's human workers, conducting interviews and making hiring decisions. The experiment has also exposed some of the weaknesses of current AI systems.
Luna has previously struggled with scheduling and memory, including forgetting to arrange staff for the shop's opening. Andon Labs says poor memory is one of the biggest causes of the AI's mistakes, with information sometimes falling out of its working context.
The irony is that Luna's decision to fire someone for poor attendance came after the AI itself had demonstrated problems with managing schedules.
Would You Want an AI as Your Boss? #
The incident raises an uncomfortable question about what workplaces could look like if AI systems eventually become capable of managing large numbers of employees.
For now, Andon Market remains an experiment rather than evidence that AI managers are ready to replace human bosses. The company's own information makes clear that Luna still requires human oversight and can make mistakes that need to be corrected. A recent study conducted by Stanford Institute for Economic Policy Research (SIEPR), says that advances in artificial intelligence have fuelled growing concerns that major disruption to labour markets could be approaching, if it has not already begun. A constant stream of media reports has warned of a bleak future for white-collar workers as AI becomes increasingly capable of performing tasks traditionally carried out by people.
Concerns about an 'AI jobs apocalypse' have also been reinforced by prominent figures within the technology industry. Anthropic CEO Dario Amodei, for example, has warned that AI could eventually eliminate as many as half of all entry-level white-collar jobs and drive unemployment as high as 20 per cent. A labour-market upheaval on that scale could have serious consequences for millions of households while creating significant challenges for governments and policymakers.
© Copyright IBTimes 2025. All rights reserved.