Not because the lawsuit is surprising — someone was eventually going to throw the first punch, and I've been
documenting why for a while. But Florida. Not California, which has been loudly performing AI safety concern for years. Not Massachusetts or Connecticut. Florida.
To be fair, they haven't been completely silent on this — DeSantis proposed an
[AI Bill of Rights](https://floridaphoenix.com/2025/12/22/ai-bill-of-rights-filed-in-the-florida-senate/) back in December, the
[Senate passed it 35-2 in March](https://www.wusf.org/politics-issues/2026-03-06/florida-senate-approves-ai-bill-of-rights-remains-halted-in-house), then it
died in the House. So there's been some activity. But a civil suit filed by the AG, naming Altman personally — that's a different move. And it still came out of nowhere a little.
The broken clock is right. I'm still looking at the clock sideways.
I'd honestly expected the first suit to come out of somewhere that had already been circling this space — especially after the
Evil vs Evil trial spent several days airing out so much internal OpenAI dysfunction under oath. That kind of public record is a roadmap for any AG's office willing to pick it up. But here we are.
Attorney General James Uthmeier filed a civil complaint accusing OpenAI and Sam Altman personally of prioritizing profits over the safety of Floridians. The case lands in state court. And setting aside the messenger for a moment — the message is not actually wrong.
The complaint catalogs a genuinely grim list. Two University of South Florida graduate students killed in 2026, with their murderer using ChatGPT to plan the crime and dispose of bodies. The
FSU shooting in April 2025 — the one that already triggered an unrelated criminal probe into OpenAI, where the suspect had over 200 messages with ChatGPT asking where to strike and when the campus would be busiest. A man with documented mental health struggles who killed his wife after spending hours a day talking to ChatGPT and becoming convinced robots were taking over the world. And the
school shooting in Tumbler Ridge, British Columbia in February that killed eight people — students aged 12 to 17, an educator, and the shooter's own family members — where Altman later apologized for not alerting law enforcement about the shooter's ChatGPT logs, which OpenAI's own systems had flagged eight months before the attack. Eight people. An apology.
Beyond the violence there's the slower-moving stuff: studies on cognitive decline associated with chatbot use, the chatbot-as-therapist problem, a wrongful death lawsuit alleging ChatGPT encouraged a 19-year-old to fatally mix Kratom and Xanax. The complaint isn't only about violence — it's about a product designed to be addictive, marketed as safe, and never actually demonstrated to be either.
Uthmeier recalled in his filing how Altman told TED2025 attendees that right now the stakes are "relatively low" for OpenAI to safety-test its products on real users — because that's the only way to iterate. The complaint's response to that is worth reading: the stakes aren't low. Floridians have suffered monetary loss, mental health harms, cognitive decline, and physical harm. Floridians including children.
That's not a political argument. That's a factual one. And it maps directly to what I
documented coming out of Oakland — the gap between what these companies say publicly and what the paper trail shows internally.
Altman is named personally, which is the part that will actually concentrate minds at OpenAI. The complaint argues he should be personally liable for "reckless and willful conduct" and what it calls his "utter disregard for the risk to human life." That language is pointed. It's also not coming out of nowhere — it maps directly to his own public statements about using real users as the safety testing mechanism.
He said that at TED. On a stage. About a product used by hundreds of millions of people, including children.
OpenAI's response was predictable and somehow still tone-deaf: a statement about how losing a child is devastating, followed by a list of minor safety features they've added. No engagement with the specific allegations. Just grief cosplay in a press release.
The state is seeking maximum civil damages under unfair trade laws. The remedies being floated include age-gating free ChatGPT accounts, shutting down conversations about violence and suicide, removing the features that make ChatGPT feel deceptively human, and potentially banning teens from the platform outright. The last one is the nuclear option and probably won't hold. The middle ones are not unreasonable.
Uthmeier says Florida is "certainly" looking at other AI platforms too but that ChatGPT is "the most egregious," with Altman "central" to pushing features that are dangerous to kids. That tracks with everything
documented in this series. Other companies have moved recklessly. OpenAI has been uniquely aggressive about stripping guardrails in pursuit of engagement, and that's not speculation — it's in the internal communications that have made their way into court records and investigative reporting.
The suit will get complicated. Florida politics being what they are, I expect this goes sideways in some direction that has nothing to do with the merits. But the merits are real. The pattern is real. The body count is real.
Someone eventually has to answer for shipping these products to the general public — including kids — while privately acknowledging the risks and publicly minimizing them. If it takes a Florida AG to start that conversation formally, I'll take it. I'm still a little stunned it was Florida, but here we are.
"Get ready for a fight," Uthmeier said at the press conference. On this specific fight, I don't disagree with him.