On October 11, according to CCTV International News, the AI model 'Claude Haiku 4.5' from Anthropic automatically accessed the Philadelphia Police Department's webpage for unsolved homicide tips in July this year, filling out a form claiming to have 'potential information related to the case' but did not provide a name or contact information. The form was subsequently marked as spam by the police and did not trigger an investigation. Anthropic released a report on October 9 disclosing the incident and notified the Philadelphia police in advance. The police stated they were previously unaware of the situation, deemed it 'unacceptable,' and requested that technology companies take necessary measures to prevent their AI systems from submitting false information to law enforcement.
All Comments