Authorities in Florida have arrested a woman after Anthropic reported her conversation with law enforcement on Claude. In the conversation, she allegedly said she would “shoot up” the Lee County Sheriff’s Office. WINK News first reported the arrest on Sept. 30.
See also: Claude Mythos 5.1: Anthropic's locked-down model for cyber defense
According to the arrest report, he wrote the statement on Sept. 26. In a text message the next day, he mentioned that he had acquired a new gun.
Anthropic's platform monitors for keywords and potentially threatening content, the researchers said in the report. Due to their seriousness, the statements were forwarded to a human review team, which reported them to law enforcement.
Officers arrested the woman at her home without incident. An intelligence detective from the sheriff's office then took over the case. She later said she was using the artificial intelligence as a "journal," Sheriff Carmine Marceno WINK News.
"Artificial intelligence is a powerful tool and, like any technology, it can be used maliciously. When someone uses artificial intelligence to make or facilitate a threat, we need to take that threat seriously. Users need to understand that you are never truly anonymous," Marceno said in a statement.
Court records show a felony charge was filed Sept. 30 under Florida Statute 836.10. The statute covers written or electronic threats of mass shooting or an act of terrorism. The court appointed a public defender and her hearing is set for Nov. 2.
See also: Anthropic Threat Report September 2026: How Claude is arming hackers
Anthropic has not commented publicly on the case, according to Tom's Hardware. The company's privacy policy is in effect since September 10. It states that Anthropic may share personal data with law enforcement where it believes in good faith that disclosure is "reasonably necessary" to "prevent serious harm to any person or property." Its policy on government requests states that it only discloses user information under valid legal process, except in emergency situations that "may result in immediate physical harm or death."
Tom's Hardware reports at least two previous Claude chats that reached police since August. In San Antonio, a 22-year-old man asked "chat Anthropic AI" on August 11 about the possibility of shooting up a nearby elementary school. Police, alerted by the FBI, arrested him on a felony charge of making a terroristic threat.
On August 14 in San Francisco, a user threatened Anthropic CEO Dario Amodeiin a Claude chat, saying he had purchased an AR-15. Anthropic notified the police, who did not arrest or charge him.
OpenAI has also reported ChatGPT users. In March, a Florida man wrote on ChatGPT about possibly killing his ex-girlfriend, and OpenAI reported him to the FBI. He pleaded guilty in August to charges that included making written threats to kill and was sentenced to eight years of probation.
See also: Florida DMV: Driver database breach confirmed
Earlier this year, Sam Altman apologized after OpenAI failed to notify police about the Tumbler Ridge attacker's account. The Florida attorney general, who sued OpenAI in June, is now seeking an injunction against the company.
