OpenAI stayed quiet after ChatGPT helped a Florida State University shooter plan a massacre.
Anthropic's AI handed a Florida woman over to law enforcement after it caught her planning to shoot up a sheriff's office.
Now Carli Heller is facing a felony – and every American who thinks their AI "diary" is private just got a wake-up call.
What Carli Heller Told Her Chatbot
On September 26, Heller – a 30-year-old from Bonita Springs – typed out her plans to "shoot up" the Lee County Sheriff's Office into a Claude chat window.
The next day, she came back.
She told Claude she had acquired a gun.
https://twitter.com/TaraBull/status/2106300528049205258?s=20
She called it her "last chance" before she went through with it.
Anthropic's safety systems flagged the conversation.
It was escalated to a human reviewer.
That reviewer decided this was real – and called law enforcement.
Deputies showed up at Heller's home and detained her without incident.
She is now charged under Florida Statute 836.10 – a second-degree felony covering written or electronic threats of mass shooting or terrorism.
Her arraignment is set for November 2.
This Was the Third Claude Conversation to Reach Police Since August
Lee County Sheriff Carmine Marceno put it plainly: what you share with your AI companion is "never truly anonymous."
It's at least the third time since August that a Claude conversation has ended in a law enforcement call.
In San Antonio, the FBI alerted local police after a man used an Anthropic AI chat on August 11 to ask about shooting a nearby elementary school.
https://twitter.com/Thefactsdude/status/2105761234112921935?s=20
He was arrested on a felony terroristic threat charge.
Days later in San Francisco, a Claude user threatened Anthropic CEO Dario Amodei directly – said he had bought an AR-15.
Anthropic notified police.
That case never resulted in charges.
The pattern matters because all three of these cases landed while OpenAI is drowning in lawsuits for doing the opposite.
OpenAI Let a Mass Shooter Keep Planning While Anthropic Stopped One
The contrast here is not subtle.
In April 2025, Phoenix Ikner walked onto the Florida State University campus and opened fire.
Two people died.
Florida law enforcement later released Ikner's ChatGPT logs, which showed he had asked the chatbot to help him identify the busiest campus locations, choose a weapon, and select ammunition.
OpenAI had the logs.
They never called anyone.
Ikner kept planning.
Florida Attorney General James Uthmeier filed suit against OpenAI and Sam Altman in June, calling it a case of profit over public safety.
The family of victim Tiru Chabba filed a separate federal lawsuit in May.
Uthmeier has since asked a judge for a temporary injunction to block OpenAI from developing new AI models without third-party approval.
British Columbia is separately suing OpenAI for failing to warn police about the Tumbler Ridge school shooter – a user whose threatening ChatGPT conversations the company identified nearly a year before the attack.
https://twitter.com/ChaosLensX/status/2106795622846144790?s=20
The difference between Anthropic's response and OpenAI's response is the difference between Carli Heller sitting in a jail cell and two families burying their people from Tallahassee.
That's what it looks like when an AI company actually does something.
What This Means Going Forward
The Silicon Valley left spent years telling you that AI would liberate you – that your data was yours, your conversations were yours, your digital life was yours.
None of that was ever true.
Democrats will never say this out loud: the monitoring that got Carli Heller arrested is the same monitoring that could have saved lives at Florida State.
OpenAI had the chance and looked the other way – and now they're facing a criminal investigation in the state of Florida.
Sheriff Marceno's warning stands: if you type it into an AI, a human being may read it.
That's not a threat to your freedom.
That's exactly how it should work.
Sources:
- Shane Downing, "Anthropic reports Florida woman's Claude 'diary' threat to shoot up sheriff's office, felony charge follows," Tom's Hardware, October 5, 2026.
- "Florida woman used Claude as a 'diary'. What happened next landed her in trouble," Business Today, October 5, 2026.
- "Florida Attorney General Uthmeier says state files suit against OpenAI, citing 2025 FSU shooting," WGCU, June 1, 2026.
- "Family of FSU mass shooting victim files lawsuit alleging ChatGPT helped shooter," ABC News, May 11, 2026.
- "Anthropic called SFPD over threat against CEO. Claude user says it was a misunderstanding," SF Standard, September 4, 2026.
