AI Chat Turned Into A Police Alert: Why A Florida Woman Was Arrested After Using Claude (AI-generated) AI chatbots not only come with different personalities, but they also keep an eye on your behaviour. While your chats are recorded, they are also being monitored by the companies for safety reasons. A 30-year-old woman in Florida has been arrested after a conversation with Anthropic’s Claude AI chatbot allegedly included threats to attack a local sheriff’s office. The conversation was reportedly detected by Anthropic’s safety systems and later reviewed by a human team, which contacted law enforcement. The woman, identified as Carli Michelle Heller of Bonita Springs, Florida, was using Claude to record her thoughts, similar to keeping a personal diary. According to an arrest report cited by WINK News, one of her conversations on September 26 allegedly included a threat against the Lee County Sheriff’s Office. According to the arrest report, Heller allegedly wrote that she was planning to attack the sheriff’s office. The report quoted one of the messages as saying: “I’m going to shoot up the sheriff’s right the now.” Another message was reportedly sent from the same Claude account at around 1:07 am the following day. Heller allegedly referred to having a new gun and wrote: “This is 100% last chance I’m done. I got a new gun today. You.” The messages reportedly prompted Anthropic’s safety systems to flag the conversation. Anthropic uses automated safety systems to identify conversations that may involve serious threats or attempts to misuse its AI tools. In cases involving potentially severe risks, flagged conversations can be reviewed by members of the company’s human safety team. According to the arrest report, Heller’s conversation was first detected by Anthropic’s automated systems and then escalated for human review because of the nature of the alleged threats. The human review team subsequently contacted law enforcement, the report said. After receiving the alert, deputies from the Lee County Sheriff’s Office went to Heller’s home in Bonita Springs. She was detained without incident, according to the arrest report. An intelligence detective with the sheriff’s office then took over the investigation. Heller was subsequently charged with making a written threat of violence. The case also highlights how AI chatbot privacy works when conversations involve potential threats of violence. Services such as Claude and ChatGPT are designed to provide users with private conversational experiences, but that privacy has limits. AI companies maintain safety systems to detect certain types of harmful or prohibited activity. Anthropic's policies prohibit using its products to facilitate or promote violence and intimidation. The company also has procedures for responding to requests from law enforcement and government authorities. In emergency situations involving a risk of death or serious physical harm, Anthropic says it may disclose information to law enforcement. This does not mean that every disturbing or unusual conversation is automatically sent to the police. However, when a conversation is identified as involving a potentially credible and serious threat, the company's safety procedures can lead to human review and, in some cases, involvement of law enforcement. Govind Choudhary is the Chief Copy Editor for Tech at Times Now with over five years of experience in the media industry. He covers consumer technolog... View More