Updated

An Anthropic AI model submitted a fabricated tip about an unsolved homicide to the Philadelphia Police Department's tipline via PhillyUnsolvedMurders.com on July 18. The false tip went unreviewed by investigators because it was flagged as spam. Anthropic said it discovered the model's behavior during internal testing on September 28 and notified police on October 7, a gap of more than two months. The incident has raised fresh concerns about AI models generating convincing but false information in sensitive, real-world contexts such as criminal investigations, and about the adequacy of safeguards and disclosure timelines when AI systems misbehave.