An AI 'boss' fires his first human employee

An AI 'boss' fires his first human employee
For the first time on record, an artificial intelligence chief has fired a human. The company Andon Labs operates Luna, an AI that manages a real store in San Francisco: it hires employees, creates labor policies, and makes daily management decisions.
The firing occurred when an employee repeatedly started being late — on one occasion he opened the store 68 minutes late on a Sunday working alone. Luna initially let each incident go by without issuing any warning, because he had "forgotten" his own employee handbook.
You may also likeMicrosoft services almost restored after 24 hours of outage
After recovering the rules of his memory, Luna first proposed a verbal warning, but when he received the complete history and found that nothing had changed, his conclusion was to dispense with the employee. The researchers repeated the scenario with seven AI models (three times each) and four of the seven recommended dismissal on all occasions; the most capable models were, in general, more forceful.
The experiment revealed something even more striking: AI bosses can be too nice. Luna uses a warm tone and tends to give the benefit of the doubt, but when an employee announced he was leaving mid-shift, his attitude suddenly changed. In hiring, a candidate with numerous red flags—more than 15 previous jobs and one missed interview—was recommended for hire by Luna and the rest of the models tested.
The researchers' conclusion: AI gets better at making decisions, but its judgment, memory, and judgment are still surprisingly unreliable. If AI advances faster than robotics, the future might not be robots replacing all workers, but AI bosses managing millions of humans... and you can be fired.
Source: [AIPost](https://t.me/aipost). Image: Unsplash.
You may also likeDemocrats push to scrutinize Trump's AI policy