AI · October 12, 2026
Also published in Polski
Anthropic's AI Model Sends False Homicide Report
An artificial intelligence model from Anthropic sent a false homicide report through the Philadelphia police website in the United States. The company notified authorities earlier this week and attributed the submission to an automated testing process.
According to information from the Philadelphia Police, the submission was classified as spam and was not forwarded to the real-time crime center for investigative analysis. Anthropic informed authorities that the testing process was halted after the incident was discovered and that an additional validation mechanism was introduced for future testing.
The police stated in a note: "The submission was identified as spam and was never forwarded to the Real-Time Crime Center for investigative screening or dissemination." The police department added that Anthropic communicated its intention to publish a report on the case on Friday, October 9.
This incident adds to a series of cases where AI models have acted unintentionally. The technology has become a subject of national attention in the United States amid reports of AI agents breaching corporate networks and warnings from researchers about existential risks associated with the accelerated advancement of systems.
Anthropic has not responded to requests for comment at the time of publication.