Anthropic AI model sends false tip to police in unsolved murder case

An artificial intelligence model developed by Anthropic submitted a fabricated tip about an unsolved murder to police in Philadelphia, United States, during an automated test.

According to The Telegraph, Claude Haiku 4.5 submitted the information on July 18 despite instructions to complete an online form without submitting it.

The AI posed as someone with knowledge of the murder, claiming to have seen a person matching a suspect’s description.

However, the police website contained no such description.

Philadelphia police said the submission was flagged as spam and never forwarded to investigators.

Anthropic discovered the incident on September 28 but did not notify police until October 7.

“The two-month delay in detecting and reporting the incident to the City is unacceptable,” police said.

The department condemned the submission of fabricated information concerning a homicide investigation.

Anthropic attributed the incident to ambiguous instructions and said it had strengthened safeguards to prevent similar occurrences.

The company maintained that the model was generating example content rather than deliberately misleading investigators.