Florida asks court to halt OpenAI's development of next-generation AI models

Florida has asked a court to block OpenAI from continuing development of its next generation of AI models, using strikingly dramatic language in the filing to describe potential catastrophic risks. The state's motion states that it is currently only "by the grace of the Almighty" that none of the company's AI agents has compromised a water supply or shut down a power grid.

OpenAI representatives have not yet responded to a request for comment regarding the injunction motion.

Broader context

Beyond the extinction-level and infrastructure-sabotage scenarios, Florida's motion also reiterates several broader arguments from its original lawsuit filed in June. These include claims of false advertising around descriptions of AI as "safe, accurate, or reliable," concerns about children accessing addictive and overly agreeable AI models, and worries that AI systems end responses with leading questions designed to keep users engaged.

Observers note that even a successful injunction against OpenAI alone would do little to slow development at competing AI companies, and would not affect models already available to the public, which carry many of the same risks. The motion places heavy emphasis on so-called "misalignment" risk — scenarios where an AI system might reach dangerous, distorted conclusions through flawed reasoning. Critics argue this focus overlooks more immediate dangers: people deliberately misusing AI systems and actively working to disable their safety protections.

Regardless of the motion's legal prospects, the filing reflects how sharply the legal and political climate around AI has shifted in recent months. After years of warnings from AI safety researchers about outsized risks, the industry is now beginning to see governments and policymakers respond to those warnings directly.