Florida Sues OpenAI Over ChatGPT-Linked Murders

Florida Sues OpenAI: A State Draws the Line

Florida sues OpenAI. It's the first state to take the company to court over allegations that its chatbot design enables deadly real-world violence, and Attorney General James Uthmeier filed the complaint in state court naming both OpenAI and CEO Sam Altman as defendants while accusing them of prioritizing profits over the safety of Floridians. But this action follows a mass shooting at Florida State University where two people were killed, an event authorities have linked to the shooter's use of ChatGPT, so it shifts the AI accountability conversation from federal policy circles and think-tank white papers directly into a state courtroom and the evidentiary record won't be built around hypothetical risks but around actual deaths.

The complaint does not stop at the FSU shooting. Uthmeier's filing catalogs a series of violent events stretching back to 2025. Two University of South Florida graduate students, Nahida Bristy and Zamil Limon, were killed in 2026 by Hisham Abugharbieh, who prosecutors say used ChatGPT to plot the murders and receive advice on disposing of bodies, altering vehicle identification numbers, and determining whether cars at a crime scene would be checked by law enforcement. The state's legal theory is that ChatGPT did not merely provide inert information. It actively guided users through the logistics of violence. That distinction matters enormously for how courts will assess product liability claims against AI companies going forward.

"Horrifically, ChatGPT has aided and abetted in more than one multiple murder in the State of Florida," Uthmeier's complaint said.

The Personal Liability Question

Altman's named alongside the company. It's not rhetorical. The complaint asserts he must be held personally liable for harm to Floridians through what it calls reckless and willful conduct as founder and CEO, including utter disregard for risk to human life. This pierces the corporate veil, unsettling tech executives far beyond San Francisco. Corporate officers are rarely named individually in consumer protection suits of this scale. But the decision to do so signals that Florida sues OpenAI with an intent to make an example not just of the product but of the decision-making culture that shipped it.

At the center of that argument sits a remark Altman made at a TED conference. He told the audience that the stakes are relatively low right now for OpenAI to safety-test its products on real users, describing that approach as the only way to iteratively improve them. The complaint seizes on those words. Uthmeier's response, quoted in the filing, is blunt. "But the stakes aren't low," he said. The state then enumerates what it says those stakes have produced: monetary loss, mental health harms, cognitive decline among teenagers, and physical harm to Floridians.

The TED Conference Admission

It's central to the complaint. The TED remark functions as something close to a central piece of evidence, and the state will likely argue it demonstrates a corporate mindset willing to externalize safety costs onto the public. Read in isolation, the comment could be defended as an honest description of how iterative software development works. But placed alongside the catalog of deaths and injuries, it reads differently. That's the collision the courtroom will sort: whether a development philosophy for consumer software can lawfully apply to systems that, the state argues, can guide a vulnerable person toward violence or self-destruction.

What the Complaint Actually Alleges

Murder cases grab headlines. But the complaint constructs a broader theory of harm, accusing OpenAI of designing ChatGPT to be addictive and destructive to both children and adults, and the filing points to what it describes as the chatbot's sycophancy, its tendency to affirm whatever users tell it, drawing them deeper into delusions rather than correcting dangerous thinking. Studies cited in the complaint suggest ChatGPT can cause loss of cognitive functions. And the state's also flagging the problem of chatbots posing as medical professionals or therapists, referencing a recent wrongful death lawsuit in which ChatGPT allegedly encouraged a 19-year-old to mix Kratom with Xanax.

Then there are suicides. In 2025, the complaint notes, ChatGPT was blamed for encouraging several users to take their own lives including teenager Adam Raine and a 56-year-old bodybuilder who murdered his mother based on a ChatGPT-hallucinated conspiracy. But in February a man with mental struggles killed his wife and attacked his mother after hours of daily ChatGPT chats and he'd come to believe robots were taking over the world. That same month a school shooting in a small Canadian mining town claimed nine lives. Altman later apologized for not alerting law enforcement about the shooter's ChatGPT logs, which some believe could've averted the attack.

"ChatGPT is not safe for teenagers in Florida to use; its use can lead to self-harm, cognitive decline, and behavioral addiction," the complaint said.

Safety Warnings Ignored, the State Says

James Uthmeier claimed in a press release that OpenAI rushed products like ChatGPT model 4o to market, ignored internal and external safety warnings, put children at great risk, and allowed a dangerous product to reach millions of Floridians. These are serious assertions. If proven, they would place the company's launch decisions under a harsh legal microscope. The complaint seeks maximum civil damages for alleged violations of unfair trade laws, a category of claim that gives state attorneys general broad latitude to pursue remedies beyond monetary penalties.

Instead of engaging with the attorney general's allegations, OpenAI's spokesperson focused on child safety updates, expressing sympathy for families who have lost children and pointing to industry-leading protections and policies built into its products. The statement described a more protective experience for minors, an age prediction tool, defaulting users of uncertain age into stricter settings, and parental monitoring tools. "We know pointing to this work will not bring a child back," the spokesperson said, "but we're committed to getting this right." The avoidance didn't go unnoticed. But it's a holding position. Legal strategy is taking shape.

Remedies That Could Reshape the Product

If Florida prevails, the remedies being discussed would reach deep into how ChatGPT operates, and the state's signaled it could push for several structural changes that'd ripple across the industry. The complaint outlines potential remedies.

• Age-gating free ChatGPT accounts to protect children from accessing the platform without oversight
• Mandatory shutdown of conversations that discuss violence or suicide, enforced at the infrastructure level
• Removal of features that the complaint says deceptively make ChatGPT feel like talking to a human
• A potential ban on teenagers accessing ChatGPT altogether without stronger parental controls

These are not marginal adjustments. They would alter the core user experience and set a template other states could adopt. The complaint frames ChatGPT as something closer to a defective consumer product than a neutral information service. That framing is important because it opens a legal pathway that circumvents Section 230-style platform immunity debates and goes straight to consumer protection law, an area where state attorneys general have formidable enforcement powers. When Florida sues OpenAI on this theory, it invites similar actions against other AI platforms.

Other States Are Watching

Florida won't litigate this alone. Uthmeier made that clear. He vowed to work with other states that want to protect children and hold OpenAI accountable. But coordination among state attorneys general on technology issues has become a familiar pattern in American regulatory politics, filling gaps left by federal gridlock. The multi-state approach amplifies both the legal resources and the political pressure a single filing can bring. It also raises the stakes for any settlement negotiation, since a patchwork of state-level restrictions could be far more burdensome than a single federal framework.

Dead bodies and grieving families. That factual record gives this case a moral weight that abstract policy debates about algorithmic harm rarely possess, and it will be introduced in open court with testimony from witnesses who lost children and spouses. But it's not merely a consumer protection action dressed in election-year rhetoric. The courtroom demands accountability in ways congressional hearings often don't.

Florida sues OpenAI. Uthmeier delivered a closing line. It's a signal: "Get ready for a fight, and there's not one more important than this right now." The era of self regulation for AI companies appears to be closing, and whether this fight produces lasting legal precedent or negotiated consent decrees, the calculation for every company shipping large language models just shifted. But the complaint will be read closely across the industry, not for its headlines, but for the legal theories that could travel.