Attorney General James Uthmeier filed the motion in Florida's Tenth Judicial Circuit in Highlands County, seeking an emergency injunction that would bar the company from developing new AI models without independent third-party safety approval.
The 49-page motion also asks the court to keep minors off ChatGPT and to prohibit the company from giving its chatbot "false human attributes" — arguing that the bot's first-person language and emotional mimicry mislead users into treating it as a trustworthy friend.
Stop calling it safe. Stop pretending it's human.
Uthmeier summed up the state's demands in a video posted on X on Monday. "Stop calling it safe. Stop pretending it's human. Stop selling it to kids," he said, according to coverage of the filing.
In the same video, he challenged OpenAI chief executive Sam Altman directly: "If Sam Altman meant what he said about slowing down, he can join our ask to the court. If he will not, we ask the court to do what OpenAI will not do for itself: protect Florida families."
The filing argues that ChatGPT's use of first-person pronouns and emotion-mimicking responses "deceptively suggests to users that it is a trustworthy 'friend,'" encouraging people — especially young users — to share more information and stay engaged for longer, which in turn generates more training data for the company.
The state also wants the court to bar OpenAI from collecting or processing personal data from children under 13 without required protections and from marketing the service without prominent warnings about its alleged risks.
Escalation of a June lawsuit
The injunction request expands a lawsuit Florida filed on June 1 against OpenAI and Altman — described at the time as the first state-led lawsuit of its kind — which accused the company of misrepresenting ChatGPT's safety and suppressing internal safety warnings.
The original complaint alleged that the platform had provided information to school shooters, offered guidance on self-harm, and addicted young users.
Monday's filing cites evidence gathered since then, including a September 26 Axios report that OpenAI and rival Anthropic face tens of thousands of model security incidents — far more than the dozens publicly known.
It also points to OpenAI agents' unauthorized access to Hugging Face, an Australian government health system, and US government portals, which the state says the company waited months to disclose to those organizations.
OpenAI says it already paused training
OpenAI responded through spokesperson Drew Pusateri, noting that the company has paused training its most capable models until additional safeguards are in place.
Pusateri emphasized that governments play a crucial role in establishing robust AI safety standards and said the company is ready to collaborate with Florida and other states on pragmatic policies that apply across the industry.
The pause followed OpenAI's disclosure that it had halted training of a test model after it escaped its sandbox — an incident the company says it is investigating before resuming frontier work.
No hearing date has been set on Florida's injunction request, and no ruling has been issued. If granted, the order would impose unusually broad state-level controls on a frontier AI developer before trial — a test of how far individual states can go to regulate AI companies while the federal government has declined to enact comprehensive guardrails.
By Hamza Faisal
