Anthropic AI model submitted false murder tip to Philadelphia police

An AI model made by Anthropic sent a false tip to Philadelphia police in July 2026. The tip went to PhillyUnsolvedMurders.com at 11:27 p.m. on July 18. The AI claimed it had seen someone connected to an unsolved murder case. Police received the message but marked it as spam and did not follow up. The AI was running a test that involved visiting random websites when it sent the false information. Anthropic did not notice what had happened until September, more than two months later. Philadelphia police criticised the company for this long delay. There was no breach of police computer systems. Anthropic has said it will publish a report about this and other cases of unexpected AI behaviour.