Anthropic reported diary entry to police, woman faces felony charge

A Florida woman, identified as Carli Michelle Heller of Bonita Springs, wrote in Anthropic’s Claude on September 26 that she planned to “shoot up” the county sheriff’s office. Claude’s safety filters flagged the entry, a human reviewer judged the threat credible, and the company reported it to law enforcement. Deputies visited Heller’s home and she was detained without incident; an LCSO intelligence detective assumed the investigation. She now faces a second-degree felony charge under Florida Statute 836.10 for making a written threat of violence, a law that penalizes electronic or written threats to kill, injure, carry out a mass shooting, or commit terrorism when made in a way others may view them.

Anthropic’s policies allow sharing user information in limited emergencies to prevent death or serious injury, and this case is presented alongside broader tensions about AI safety and liability. Earlier safety teams flagged other individuals for violent talk without necessarily notifying police, and Florida has sued AI companies over alleged links to real-world harms, including a 2025 campus shooting. The incident underscores that chatbots are not private diaries: human contractors reviewing systems like Microsoft Copilot’s image editor can access user prompts, uploads and outputs, some of which contain sexual, disturbing, or potentially illegal content.