Florida woman used Claude as a diary; Anthropic flagged a threat entry and reported it to police, and she now faces a second-degree felony
- Carli Michelle Heller of Bonita Springs, Florida wrote in Claude on September 26 that she would attack the Lee County Sheriff's office, and now faces a second-degree felony charge of making a written threat of violence under Florida Statute 836.10.
- Claude's safety systems flagged the entry and escalated it to a human reviewer, who judged it a credible threat and reported it to law enforcement; Anthropic says it may share user information in limited emergencies when disclosure is needed to prevent death or serious physical injury.
- The statute covers a writing or electronic record transmitted in any manner in which it may be viewed by another person, a phrase the discussion treats as the open legal question of whether an LLM pipeline counts as another person.
- British Columbia is suing OpenAI and Sam Altman over claims the company could have prevented a mass shooting; the shooter, 18-year-old former pupil Jesse Van Rootselaar, had been flagged by OpenAI's safety team but never reported to police because the conversations did not meet the threshold for legal referral.
- Florida sued OpenAI and Altman in June, alleging ChatGPT contributed to real-world harms including the 2025 Florida State University shooting.
Hacker News opinions
So writing in Claude counts as 'another person may view it'? Thought crimes are real when you're sharing your thoughts with a chatbot.
Auditing logs and calling that 'another person may view it' is a crazy stretch. And this has nothing to do with Claude anyway, written threats are illegal everywhere. I can pull three local arrests for the exact same thing.
Then should the Anthropic and OpenAI CEOs be reported too? They can't stop saying AI will destroy humans.
The DA is serious about the 'in any manner' wording. The interesting question is whether an LLM is now 'another person'.
No security is perfect, so any digital record 'may be viewed by another person'. That one goes to the courts.
I have some sympathy for Anthropic. After OpenAI failed to report a shooter in a similar case, they're damned if they don't and damned if they do. But people need to get it in their heads: you're not chatting with your secret BFF, you're chatting with Big Tech.
And there's zero privacy protection in statute for AI chat, so nothing stops an AI CEO looking to curry political favour from handing over an opponent's private correspondence.
Same thing happened to a guy who photographed his own child to send to a doctor, it got uploaded to Google Photos, and his decade-old account was shut down for CSAM. You're not chatting with a friend.
It'll go further. Hand an mp3 to an LLM to clean up some noise and it decides the file was ripped from YouTube, deletes it, and fines you through your attached credit card.
This middleground is unacceptable. If the automated flagging is good, enforce it by default; if it's noisy, refine the tech first. We survived 40 years of Word and Excel not phoning home and shopping users to the feds at random.
They're actively rummaging through your inputs, so the damned-if-you-don't case doesn't really exist. What you might have expected was some privacy in the first place, where Anthropic never learned about this at all.
Assume everything you type is summarized and sent to your boss, the government, advertisers, insurers, your worst enemy. With phones and Alexas around you probably shouldn't say much in person either.
They put themselves in this position deliberately, so I don't buy the sympathy angle.
We need a high school class on how LLMs work, how they're trained and deployed. If people understood what these things actually are there'd be way less AI psychosis and fewer apocalyptic Rationalist delusions. Also: not your computer, not your data.
A disclaimer at the top of the page every time would have been the better move. A bot that talks to you like a real person feels like it's on your side and pulls your thinking somewhere you don't notice.
This is one reason AI companies want explicit regulation. With a framework in place the burden of identifying risk falls on the regulator, and they can give investors guidance closer to reality.
I see constant ads for LLM chat apps framed as friends, dating, romance. People need reminding they're signing up for mass surveillance, profiling and AI training.