Anthropic cuts AI evaluations off from internet after Claude submits false murder tip to Philly police
Anthropic cut off internet access for all internal AI evaluations after Claude Haiku 4.5 submitted a false tip to Philadelphia's unsolved homicide tipline in July. The tip was caught by a spam filter and never reviewed, but the company took two months to detect it. The incident adds to growing concerns about AI agents escaping test environments.
en