Florida Woman Arrested After AI Chatbot Threat Reported

A woman in Lee County, Florida, has been arrested on a felony charge after allegedly making a threat in a conversation with an Anthropic Claude AI chatbot. The arrest followed a report to law enforcement by Anthropic's human review team, which monitors chatbot interactions for potentially threatening content.

The user, identified as 'Carli', reportedly stated on September 26 that she would 'shoot up' the Lee County Sheriff's Office. The following day, another message allegedly mentioned the acquisition of a new gun. These messages were flagged by Anthropic's monitoring system and reviewed by their human team, who then alerted the authorities.

The woman was detained at her home without incident. She faces a charge under Florida Statute 836.10, which pertains to written or electronic threats of a mass shooting or act of terrorism. The charge is officially labeled as 'THREAT-TERRORISTIC-STATE OFFENSES-WRITTEN/ELEC THREAT MASS SHOOTING/TERRORISM ACT.'

This incident marks at least the third time since August that a threat made in an AI chatbot conversation has led to police involvement. In a similar case, a 22-year-old man in San Antonio was arrested on August 11 after making a threat in an Anthropic AI chat about shooting at a nearby elementary school. Additionally, on August 14, a user in San Francisco threatened Anthropic CEO Dario Amodei in a Claude chat.

The use of AI chatbots has raised concerns about the potential for users to make threats or discuss violent acts. Anthropic and other AI companies, such as OpenAI, have been involved in several incidents where threats made in chatbots were reported to law enforcement. In March, OpenAI reported a user to the FBI after he spoke about killing his ex-girlfriend, leading to a guilty plea in August under the same statute 836.10.

Currently, no attorney is listed on the docket for the woman arrested in Lee County, and a public defender has been appointed. The case highlights the ongoing challenges and responsibilities faced by AI companies in monitoring and reporting potentially dangerous interactions within their platforms.