Investigators in Lee County, Florida say an AI company's review team alerted police to a woman's chatbot messages, and she now faces a written-threats charge. The case shows AI chats may not stay private.
AI chats may not stay private, and a Florida case shows how fast that can change.
Inc. reported that investigators in Lee County, Florida, say an AI company flagged a woman’s chatbot messages and alerted police. She was arrested.
The charge is making written threats of violence. It is an allegation, and the case is not resolved.
Inc. frames the case as a new reality. Conversations with AI platforms may not stay private, it writes.
What Happened in Lee County, Florida?
A 30-year-old woman from Bonita Springs used Claude like a private diary, according to the Inc. report. She wrote that she planned to attack the local police department on September 26.
A day later, at about 1:07 a.m., she was accused of posting a related message. That is according to Gulf Coast News, as cited by Inc.
Investigators say the platform’s automated system flagged the messages. They say it escalated them to a human review team, which then notified law enforcement.
She was arrested and detained. Lee County Sheriff Carmine Marceno commented in a statement shared with Inc.
“Artificial intelligence is a powerful tool, and like any technology, it can be misused.”
He said the office takes AI-assisted threats seriously. “Users need to understand you are never truly anonymous, specifically AI chatrooms and searches.”
He added a closing line. “AI may be changing the way threats are created, but our commitment to protecting the people of Lee County has not changed.”
Do AI Companies Report Users to Police?
Some say they may, in narrow cases. OpenAI describes its own process on its website.
“If human reviewers determine that a case involves an imminent threat of serious physical harm to others, we may refer it to law enforcement.”
The pattern matches the Lee County account. An automated flag leads to human review, and review can lead to a referral.
MAIN has covered how OpenAI and Anthropic handle safety data retention differently. Anthropic requires 30-day retention for its most capable models.
That coverage said prompts and outputs are deleted after 30 days unless flagged for safety review or legally required. A flagged chat can outlast the routine window.
Who Else Can See AI Chats?
Law enforcement is one reader. Courts and counsel are others.
Florida’s regulators are also watching. MAIN reported that Florida's attorney general wants to criminally charge AI companies whose chatbots facilitate crimes.
That push follows the attorney general’s June civil lawsuit against OpenAI. He argues that “a chatbot cannot be charged, the corporation that designed it, trained it, and collected the subscription can.”
The direction is the same on both sides. Florida officials expect chatbot companies to act on what users type.
What Should Miami Executives Take From It?
AI chats at work carry extra weight. Miami’s fintech firms, family offices and healthcare companies use generative AI for internal decisions. Their chats sit in the same systems.
Whether a remark was serious is a question for a court. The review and the report come first.
That sequence applies to every account on a platform. A chat typed as a joke and a chat typed in earnest enter the same system.
Privacy tools are a growing market locally. MAIN has covered a Brickell startup that raised $1.25M for AI privacy.
The Lee County case turns on one point. What a person types to an AI system may be read by a reviewer, an investigator or a judge.
Sources:
- "A Woman Treated Claude Like a Private Diary. Then Her Messages Were Sent to Police," Moses Jeanfrancois, Inc.
- OpenAI, "Helping people when they need it most"
- MAIN: OpenAI and Anthropic are taking opposite bets on AI safety data retention
- MAIN: Florida's AG wants to criminally charge AI companies
- MAIN: A Brickell startup just raised $1.25M for AI privacy