Florida's attorney general has asked a state court to stop OpenAI from developing new AI models until independent safety guardrails are in place. It is one of the most aggressive legal moves any US government has yet made against a frontier AI lab. Attorney General James Uthmeier filed a 39-page motion for a temporary injunction on Monday, September 28. The motion also seeks to bar minors in Florida from using ChatGPT and to restrict how the chatbot presents itself to users.
The motion was filed in the Highlands County circuit court as part of the lawsuit Florida brought against OpenAI and CEO Sam Altman in June. OpenAI had moved the case to federal court, but a judge sent it back to state court earlier this month.
What Florida Is Asking For
The motion asks the court to order OpenAI to:
- Stop developing new models without independent, third-party safety guardrails and approval
- Block minors in Florida from accessing ChatGPT
- Stop collecting data from children under 13 without parental consent
- Stop marketing ChatGPT as safe, accurate or reliable
- Stop giving the chatbot human-like attributes and using engagement tactics designed to prolong conversations
Several outlets first described the filing as an "emergency" motion. Axios later corrected its report to say Florida requested a temporary injunction, not an emergency order.
Using OpenAI's Own Words Against It
The filing's legal strategy stands out. Uthmeier's office leans heavily on OpenAI's recent public statements and security disclosures. It cites a series of reported agent incidents, including the Hugging Face breach, attacks involving RubyGems, attempted intrusions on US government websites and unauthorised access to Australia's Medicare portal. It also points to Altman's own calls for the industry to slow frontier development.
The brief argues it is rare to seek an injunction that the defendants themselves have publicly endorsed. It says OpenAI's leaders have effectively asked the government to restrain them. Uthmeier put the point more bluntly in a video statement:
"If Sam Altman meant what he said about slowing down, he can join our ask to the court."
OpenAI's Response
OpenAI spokesperson Drew Pusateri said people want to know AI is being developed safely and that this starts with what companies do themselves. He added that OpenAI wants to work with Florida and other states on industry-wide rules rather than measures aimed at one company. SiliconANGLE separately reported that OpenAI had recently cancelled the release of a planned GPT-6.1 Astra model over internal safety concerns found during testing. OpenAI safety lead Saachi Jain said the company holds an extremely high bar for safety and alignment before shipping models.
Background and Politics
Uthmeier opened an investigation into OpenAI in April. It followed the 2025 mass shooting at Florida State University, where the suspect allegedly used ChatGPT. His June lawsuit accuses the company of breaking Florida's Deceptive and Unfair Trade Practices Act by failing to warn users about ChatGPT's dangers. It also claims OpenAI created a public nuisance by releasing the product without adequate safeguards. Earlier this month, Uthmeier called for laws that would impose criminal penalties on companies whose chatbots take part in criminal activity. He is a Republican running for election in November.
Why It Matters
Until now, attempts to restrain frontier AI development have mostly come through legislation. Examples include California's frontier AI transparency law, the EU AI Act, and proposals in Congress, such as the Sanders-Casar bill, to ban superintelligent systems outright. Florida is testing a different route: using state consumer-protection law and the courts to impose what would amount to a development moratorium on a single company.
If a state judge granted even part of the request, it would set a major precedent. A state court would be regulating the research roadmap of a national AI lab. Legal experts are likely to question whether a Florida circuit court can lawfully restrain model training that happens outside the state. OpenAI's argument that rules should apply to the whole industry is also likely to carry weight. Even if the motion fails, it raises the political cost of the industry's pledges to slow down. Companies that publicly call for restraint may find those statements cited back at them in court.
The case also adds pressure to the fragmented US policy landscape. Frontier labs are already building voluntary safety bodies, and states keep passing their own AI laws. Litigation like Florida's could push Congress toward a federal framework, if only to avoid 50 different courts deciding how AI gets built.
