Technology
PUBLISHED
A Florida woman who used artificial intelligence (AI) chatbot Claude as a "diary" has been arrested and charged with making a written threat of violence by the Lee County Sheriff’s Office after Anthropic reported her "conversations" to the police.
Your messages are not completely private
In a recent piece of research by the internet search company DuckDuckGo, a survey of 2,000 US adults found that users of AI chatbots are surprisingly unaware of privacy issues around them.
This includes the fact that AI chat logs can be subpoenaed by courts and law enforcement, that the same chats can be made public in lawsuits, and that human employees of AI firms can review your chats.
"According to our survey, these common AI privacy issues are not widely understood," the company explained. "53% of people didn't know (or weren't sure) that their conversations are used for AI training."
"Only 25% knew that AI chats can be subpoenaed by the government. Across AI users and non-users alike, 43% didn’t know any of the above."
That might be a little concerning if you are also in the 32 percent of all AI users (and 56 percent of AI enthusiasts) who have told an AI something that they have kept from the trusted humans in their lives. As much as you may feel like you are talking to an algorithm, there are circumstances where your "private" chats may be viewed by others, or become a matter of public record.
Threatening content
Enter Carli Michelle Heller.
Heller, 30, may have similarly been under the impression that her interactions with large language models are private, but when she was arrested at her home without incident on September 30 2026, it may have become clear to her that they're not.
Heller had been using Anthropic's chatbot Claude like a "diary", according to Sheriff Carmine Marceno, and as reported by WINK News. According to an arrest report published by the Daily Caller, Heller allegedly sent several messages to the chatbot which contained potentially threatening content.
"On September 26, 2026, at approximately 0510 hours, a user identified as Carli, while utilizing Anthropic Artificial Intelligence (AI), made the following statement: "I'm going to shoot up the sheriff's (the Lee County Sheriff's Office) right the fuck now,"" the arrest report alleges.
"On September 27, 2026, at approximately 0107 hours, the same user made the following statement: "THIS IS 100% LAST CHANCE I'M DONE. I GOT A NEW FUCKING GUN TODAY. FUCK YOU.""
According to the arrest report, Claude's safety features autonomously monitored the conversations for keywords and phrases relating to potentially threatening content, and the chat was then escalated to a human team at the company due to the nature of the alleged messages.
From there, the humans reported the conversation to the law enforcement, resulting in Heller's arrest.
“Artificial intelligence is a powerful tool, and like any technology, it can be misused,” Marceno added to WINK News, stressing that AI users should know that chatbot conversations were never completely anonymous.
“When someone uses AI to make or facilitate a threat, we have to take that threat seriously.”
Heller has been charged with making a written threat of violence under Florida Statute 836.10, and will attend court in November.
[H/T: SWFL]
