Anthropic Reported User's Claude 'Diary' to Police
Original: Anthropic reported diary entry to police, woman faces felony charge
Why This Matters
Sets a precedent for when AI companies are legally and ethically obligated to report users.
Florida woman Carli Michelle Heller faces a second-degree felony after Anthropic's Claude flagged her Sept. 26 entry threatening to 'shoot up' a Sheriff's office. A human reviewer escalated it to law enforcement.
Carli Michelle Heller of Bonita Springs, Florida, now faces a felony charge after writing a threat against a local Sheriff's office in Claude—a chatbot she reportedly used as a personal diary. The entry, dated September 26, was flagged by Claude's safety systems and escalated to a human reviewer who deemed it credible enough to report to police. Deputies visited Heller's home and detained her without incident before an LCSO intelligence detective took over. Anthropic's privacy policy permits sharing user data in limited emergencies when disclosure may prevent death or serious physical injury. Heller is charged under Florida Statute 836.10, which makes it a second-degree felony to transmit a written or electronic threat of violence, mass shooting, or terrorism in a manner viewable by others. The case arrives amid broader scrutiny of AI companies' roles in violence prevention. OpenAI faces a lawsuit from British Columbia alleging it failed to act on flagged conversations by a shooter who later killed people, and Florida separately sued OpenAI over ChatGPT's alleged connection to the 2025 FSU shooting. The contrast is stark: OpenAI said prior conversations didn't meet its threshold for referral; Anthropic's did.