Florida Woman Used Claude as a Diary. Anthropic Reported Her Threat to the Police.
Trending Topics Jakob Steinschaden
A Florida woman told Claude she’d “shoot up” a sheriff’s office, and Anthropic passed it to police. It’s a blunt reminder that chatbot chats aren’t private diaries.
Based on reporting by Trending Topics, Jakob Steinschaden — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
A 30-year-old woman from Bonita Springs, Florida, is facing a second-degree felony after Anthropic flagged violent threats she made to Claude and sent them to law enforcement. The arrest report says she had been using the chatbot like a personal diary. Then she crossed the line from venting into explicit violence.
According to Tom’s Hardware, the first alarming message came in late September, when she wrote that she was going to “shoot up” the Lee County Sheriff’s Office. A day later, she said she had bought a new gun. Anthropic’s automated systems caught the messages, a human reviewer judged them to be a credible threat, and the company contacted the authorities. Deputies arrested her at home a few days later without incident.
Florida law treats written or electronic threats to kill, carry out a mass shooting, or commit terrorism as a crime under Section 836.10. A public defender has been appointed, and she is presumed innocent. The legal machinery here is plain enough; the uncomfortable part is how ordinary the path was from chatbot confession to handcuffs.
This appears to be at least the third Claude conversation reported to police since August, again according to Tom’s Hardware. In San Antonio, a 22-year-old was arrested after asking about a school shooting. In San Francisco, a user threatened Anthropic chief executive Dario Amodei and bought an AR-15-style rifle, though no arrest followed. And that is exactly the problem tech companies keep running into: people want a private therapist with a server farm behind it, but the servers keep filing incident reports.
Anthropic says its privacy policy allows disclosure when it believes, in good faith, that it is needed to prevent serious harm, and it has said those disclosures are limited to emergencies involving a risk of death or serious physical injury. OpenAI is dealing with the same mess from the other side, with lawsuits and its own threat reporting. The bigger lesson is brutally simple: these chatbots are not locked journals, no matter how much users wish they were.
My take — AI-written commentary, not fact-checked reporting
The industry keeps selling intimacy and then acting surprised when intimacy brings liability. Chatbots are becoming the new confessional, except the priest is a moderation system and the curtain can’t be trusted. Regulators should stop pretending this is normal software and force clearer rules before every “private” chat becomes a police report.
Read more about this at: Trending Topics