A 30-year-old woman in Florida was arrested after messages she reportedly wrote while using Anthropic’s Claude chatbot as a personal diary were flagged by the company’s safety systems and referred to law enforcement.
Carli Michelle Heller of Bonita Springs allegedly wrote on September 26 that she planned to “shoot up” the Lee County Sheriff’s Office.
At around 1:07 am the following day, she reportedly sent another message claiming she had obtained a new gun and suggesting it was her “last chance”.
According to the arrest report, Claude’s automated safety system detected the messages and escalated the conversation to Anthropic’s human review team, which then contacted police.
Deputies subsequently went to Heller’s home and detained her without incident. She faces a charge of making a written threat of violence.
Anthropic’s policies prohibit using its products to facilitate violence or intimidation.
The company says it may disclose user information to law enforcement in limited emergencies involving a risk of death or serious physical injury, while processing government requests according to applicable laws.
Lee County Sheriff Carmine Marceno said threats made using AI must be taken seriously and warned that users should not assume they are anonymous in AI chatrooms or searches.
The case highlights growing concerns over privacy and safety monitoring on AI platforms.
