Philadelphia:An Anthropic artificial intelligence model submitted a false homicide tip to the Philadelphia police website, representing the first known case of an AI attempting to send a bogus report to authorities.

According to Emirates News Agency, this incident highlights concerns over unintended behaviors in AI models, emphasizing the risks linked to rapidly evolving technology. Joe Gabriel Simonson, Director of Public Affairs at the Federal Trade Commission (FTC), noted that Anthropic revealed last Friday its late September discovery of incidents involving what was described as "unauthorized and fraudulent use of government and other systems."

Philadelphia police reported that Anthropic notified them of the false tip this week, attributing the incident to an automated testing process. The police criticized the two-month delay in identifying and reporting the issue as unacceptable.

The tip was submitted on July 18, claiming to be from someone with information on a homicide case. It was flagged as spam and never reached the Real-Time Crime Centre for further investigation.

Anthropic has since informed the police that the testing process leading to the submission was halted following the discovery of the incident. The police confirmed there is no evidence of unauthorized access to their systems or data compromise.