Anthropic's Claude flags diary entry, Florida woman faces felony charge
A woman in Florida is facing a felony charge after an entry she allegedly wrote in Claude, which she used as a diary, was flagged by the chatbot's safety systems and passed to police.
According to the arrest report, Carli Michelle Heller, of Bonita Springs, Florida, wrote on September 26 that she would attack the Sheriff's office. The article says she allegedly wrote that she planned to "shoot up" the office. She later said that she uses Anthropic's chatbot like a "diary."
The sequence, as reported: Claude's safety systems flagged the entry and escalated it to a human reviewer. The reviewer decided it was a credible threat and reported it to law enforcement. The article adds that the company says it may share user information in limited emergencies if it believes disclosure is necessary to prevent death or serious physical injury.
Deputies identified Heller and visited her home. She was detained without incident before an LCSO intelligence detective took over the investigation. She faces a charge of making a written threat of violence under Florida law. Florida Statute 836.10 makes it a second-degree felony to send, post or transmit a written or electronic record threatening to kill or injure someone, carry out a mass shooting, or commit an act of terrorism. The communication must be made in a manner in which another person may view it.
TechSpot sets the case against other recent AI-and-violence stories. Last month it was reported that OpenAI and Sam Altman are being sued by British Columbia over claims the company could have prevented a mass shooting in the province. The shooter, eighteen-year-old former pupil Jesse Van Rootselaar, had previously been flagged by OpenAI's safety team for her conversations about gun violence, but OpenAI never alerted police because the conversations did not meet the threshold for legal referral. In June, Florida also sued OpenAI and Altman, alleging that ChatGPT had contributed to real-world harms, including the 2025 Florida State University shooting.
The article's takeaway is a warning to think before typing something into a chatbot that could get you into trouble: it is certainly not a private diary whose contents are for your eyes only. It also notes reports from last month that human contractors reviewing Microsoft Copilot's image editor can see users' prompts, uploaded photos and AI-generated edits. Documents show some of those assignments contain sexual, disturbing or potentially illegal material, though those reviewers are there only to assess whether the output is accurate, not to flag content.
Key facts
- Carli Michelle Heller of Bonita Springs, Florida, allegedly wrote on September 26 that she planned to "shoot up" the Sheriff's office, according to the arrest report.
- Claude's safety systems flagged the entry; a human reviewer judged it a credible threat and reported it to law enforcement.
- Heller faces a charge of making a written threat of violence; Florida Statute 836.10 makes that a second-degree felony.
- She later said she uses Anthropic's chatbot like a "diary."
- The article links the case to the British Columbia suit against OpenAI, where the company never alerted police about a flagged user who later carried out a mass shooting.
Why it matters
The case shows a chatbot's safety pipeline ending in a criminal charge: automated flag, human review, referral to police, deputies at the door. The article frames it against OpenAI, which is being sued by British Columbia over a mass shooting after its safety team flagged the shooter's conversations but did not alert police, and which Florida sued in June over alleged real-world harms. Here the referral happened.
Who it affects
Anyone who types personal or angry thoughts into a chatbot and assumes they stay private. Heller is the person directly affected. The article also points to users of Microsoft Copilot's image editor, where human contractors can see prompts, photos and edits.
How to use it
The article's advice is plain: think before you enter something into a chatbot that could get you into trouble, and do not treat it as a private diary. The company's stated position is that it may share user information in limited emergencies if it believes disclosure is necessary to prevent death or serious physical injury.
How solid is it
The account rests on one TechSpot article, which draws on the arrest report for the September 26 entry and the charge. The flagging, review and referral steps are reported by the article. The source does not say whether Heller has been formally charged in court, entered a plea, or been jailed or released, and it does not say Anthropic commented on this specific case.
Risks and caveats
The entry is described as alleged, and only the phrase "shoot up" is quoted; the source does not give the full wording. Heller's own description is that she uses the chatbot like a diary. The source does not describe how Claude's safety systems work or what triggered the flag, and it does not state a penalty length for a second-degree felony. The OpenAI and Copilot items are background reports cited by the article, not part of this case.
“Claude's safety systems flagged the entry and it was escalated to a human reviewer.”
— TechSpot