In the US, artificial intelligence fired an employee who was late for work

In the US, artificial intelligence fired an employee for the first time. This was reported by Time magazine.

An artificial intelligence assistant named Claude was hired to manage a store in San Francisco. The startup Andon Labs introduced an artificial intelligence model into the store's operations in order to study the capabilities of neural networks and their impact on the economy. The store employees were people with real employment contracts.

Claude, acting as a manager, fired one of his subordinates. The reason for the dismissal was that the employee was late for 17 out of 23 shifts. Initially, the artificial intelligence had lost its memory of the instructions it had developed. Therefore, as a manager, it "entertained" and taught the late employees not to worry.

The neural network was in constant contact with an employee of Andon Labs, asking him to intervene in the situation. Following the instructions given by the person, Claude found a forgotten manual and detected regular delays. Initially, the assistant recommended issuing a formal warning, but later decided to dismiss.

From a financial point of view, the experiment did not pay off. In five months, the store's balance sheet decreased from $ 100 thousand to $ 61 thousand. Claude's indecisive management decisions and questionable business acumen were the cause of this loss.

Earlier, it was revealed that the acting head of the US Cybersecurity Agency had leaked service files to ChatGPT. These files consisted of CISA contract documents marked "for use only within the service" and not intended for public distribution.