Sam Altman has publicly called for the AI industry to slow down. Now the Florida attorney general is asking a court to hold him to that. In a motion filed Monday, AG James Uthmeier requested an emergency injunction that would bar OpenAI from training new models without independent oversight and cut minors off from ChatGPT entirely. The filing is the most aggressive legal escalation yet in what has become a sustained state-level campaign against the company.
The backstory starts with the 2025 mass shooting at Florida State University. The suspect allegedly used ChatGPT in the lead-up to the attack. Florida launched a criminal investigation into OpenAI in April of that year, then filed a formal lawsuit in June. The emergency motion filed this week is the next step, and it is direct. “No new model development without independent safety guardrails; no more harvesting children’s data; no more calling this product safe, accurate, or reliable,” Uthmeier said in a video posted to X. “If Sam Altman meant what he said about slowing down, he can join our ask to the court.”
OpenAI has consistently pushed back on the central claim. After the initial criminal investigation launched, the company told Engadget that “ChatGPT is not responsible for this terrible crime” and described the tool as a general-purpose product used by hundreds of millions of people for legitimate purposes. The company did not respond to requests for comment on the new motion.
What makes the injunction request strategically interesting is how Uthmeier’s office built the justification. They cite OpenAI’s own public statements, including Altman’s calls for the industry to slow frontier model development, and the company’s recent security disclosures tied to the Hugging Face hack. The brief states bluntly: “It is a rare request for an injunction where the Defendants themselves have publicly endorsed it.” That framing is either clever lawyering or political theater, depending on your read. Probably both.
This is also not the first time OpenAI has been pulled into the orbit of a mass violence case. In February, Canadian regulators pushed the company to revise how it handles threats of harm, following a Wall Street Journal report that OpenAI had flagged the account of an alleged shooter in 2025 but did not contact law enforcement. OpenAI later agreed to updated policies for working with Canadian authorities.
For the AI industry broadly, the Florida case is worth watching closely. State-level legal pressure is a different beast from federal regulatory proposals. If a court grants even a partial injunction, it sets a precedent that competitors like Anthropic, Google DeepMind, and Meta would have to take seriously. The legal theory being tested here, that an AI company bears liability for how its tool is used in violent crimes, has not been tested at this scale. And with elections approaching and AI safety still a politically charged topic, Uthmeier’s move is almost certainly the start of something, not the end.
