Anthropic Reports Florida AI User to Police Over Written Sheriff Threat

Carli Michelle Heller of Bonita Springs, Florida, was arrested and faces a felony charge after Anthropic reported threats made through her Claude account to law enforcement, according to Decrypt and Tom's Hardware. Decrypt reported that Heller had been using the artificial intelligence platform as a personal diary before the interaction was reviewed. [1][3]

Local outlet WINK News, cited by Decrypt, reported that Heller wrote in her Claude account on September 26 that she would shoot up the Lee County Sheriff's Office. Investigation records showed that her conversations on the platform indicated she had acquired a new firearm by the following day, Decrypt reported. [1]

Decrypt reported that Anthropic uses automated software to monitor user interactions for key phrases and threatening content. After Heller's diary entries triggered the safety filters, a human review team reviewed the messages and sent the information to police investigators, according to Decrypt. [1]

Anthropic's terms of service allow the company to report user communications to law enforcement when severe safety risks are detected, Decrypt reported. Tom's Hardware reported that this case marks at least the third time since August that a conversation with an artificial intelligence chatbot has been turned over to police. [1][3]