An Anthropic AI model reportedly submitted a false homicide tip through the Philadelphia Police Department’s website during testing. Authorities said the submission was flagged as spam and did not lead to an investigation, per The New York Times.