A 30-year-old woman from Lee County, Florida, was arrested after using the Claude chatbot as a personal diary. The suspect, Carly Michelle Helle, allegedly wrote threats of violence against the Lee County Sheriff's Office within the chat interface. Anthropic's safety systems flagged the messages, which were subsequently reviewed by human moderators who notified law enforcement.
According to the arrest report, Helle sent messages on September 26 stating she intended to shoot at the sheriff's office. The following day, she claimed to have acquired a new firearm. Police arrived at her residence in Bonita Springs and detained her without incident. She currently faces a charge of written threats of violence, though she has not yet been convicted.
Anthropic's policy allows the company to disclose user information to authorities without standard legal processes if there is an imminent risk of physical harm or death. The company employs a combination of automated detection and human review to enforce its safety policies. This incident highlights the privacy limitations of AI platforms, as user inputs are subject to monitoring.
The case has sparked broader discussions regarding the legal boundaries of AI privacy and the responsibilities of tech companies. Legal experts note that while prosecutors may view the messages as direct threats, the case involves complex arguments regarding free speech. Meanwhile, other AI firms like OpenAI are currently facing lawsuits related to their roles in preventing real-world violence.
