AI Firms Report Chats to Police

Anthropic’s human reviewers flagged a Claude chat in which a user allegedly threatened a sheriff’s office and reported it to law enforcement, leading to a felony arrest under a Florida statute. The incident shows how AI companies’ monitoring and disclosure policies can trigger criminal investigations even absent formal government requests. — This matters because platform‑level monitoring plus emergency‑disclosure rules effectively deputize private firms as first responders and evidence‑gatherers, raising questions about privacy, due process, transparency, and who decides when speech merits police involvement.