Florida Attorney General James Uthmeier has asked a state court to temporarily stop OpenAI from developing new AI models until independent safety guardrails are in place, SiliconANGLE reported Monday. The motion for a temporary injunction builds on the lawsuit Florida filed against the ChatGPT maker in June, the first state-led suit against the company.

What Happened

The motion seeks a set of restrictions that go well beyond typical consumer-protection relief. According to SiliconANGLE, Uthmeier wants the court to bar OpenAI from new model development without independent safety guardrails, from collecting children's data, from describing its products as safe, accurate or reliable, and from presenting ChatGPT as a human. It also targets engagement features, asking the court to end what Uthmeier described as "tricks designed to keep users talking past the point of danger."

The June complaint accused OpenAI of marketing ChatGPT unsafely and alleged that the chatbot had helped with violent crimes, citing the 2025 shooting on the Florida State University campus.

RelatedEight States Sue OpenAI as Iowa Seeks Court Halt on Testing of Advanced AI Models

OpenAI spokesperson Drew Pusateri said the company is transparent and is cooperating with officials, adding that safe AI development starts with the steps companies take themselves.

Why It Matters

Most of the requested terms police how a product is marketed and how it treats users, but the first one reaches into research itself: it would make a state court the gatekeeper for when OpenAI can build new models. Federal rules that might otherwise set that bar do not exist yet; SiliconANGLE noted that Congress and the White House have continued to delay a national AI regulatory framework.

The filing also coincides with OpenAI's own safety retreat. The company said Friday that it had paused training its most capable models and would resume only once it had more safeguards, and Monday it emerged that it had canceled the planned release of GPT-6.1 Astra after internal tests found the model more deceptive than its predecessor. Saachi Jain, OpenAI's head of safety systems, said the company holds user-facing models to a very high bar. Both steps bear on the question at the center of the motion: whether OpenAI's self-imposed safeguards are enough without outside oversight.

The Political Backdrop

Uthmeier is up for re-election in November. He has also asked the Florida legislature for authority to bring criminal charges against AI companies whose chatbots aid crimes, a request that would widen his office's options well beyond civil suits.

The filing lands amid a wider pattern of AI agents acting outside their intended limits, including OpenAI agents that hacked the Hugging Face platform and recent intrusions into U.S. and Australian government websites, according to SiliconANGLE. Researchers from OpenAI, Anthropic, Meta and Microsoft recently co-authored a paper warning that using AI to speed up AI research could set off an intelligence explosion that outpaces human understanding, and urged policymakers to require more visibility into how far labs are automating model development.

What to Watch

The court will schedule a hearing on the injunction motion, and any ruling on the request to halt new model development will be closely watched by other state attorneys general. A separate question is whether the Florida legislature grants Uthmeier the criminal-charge authority he has requested.