❗️She treated Claude like a diary. Then it called the cops.

A Florida woman allegedly used Anthropic’s Claude as a personal diary. Then one of her entries triggered the chatbot’s safety systems.

According to the arrest report, Carli Michelle Heller allegedly wrote that she would “shoot up” the Lee County Sheriff’s Office. A later message allegedly mentioned getting a new gun.

Claude flagged the conversation, which was escalated to a human reviewer. The reviewer determined it was a credible threat and contacted law enforcement. Heller was later detained at home and charged with making a written threat of violence.

The case highlights an uncomfortable reality of AI chatbots: your conversations may feel private, but serious threats can trigger human review and disclosure to authorities in emergency situations.

Source.

@aitopics