Florida sues OpenAI and Altman over ChatGPT safety

Florida on Monday became the first US state to sue OpenAI and chief executive Sam Altman, accusing the ChatGPT maker of promoting its chatbot as safe while overlooking warnings about harm to users and children. The case, announced by Florida attorney general James Uthmeier, puts a live legal test around safety claims made by the company behind ChatGPT.

Florida says OpenAI made broad safety assurances while expanding ChatGPT to a mass audience. It wants a court to examine whether those statements amounted to deceptive practices. The allegations have not been tested, and OpenAI has signalled it will defend its safeguards.

Uthmeier said the case names both the company and Altman because, in Florida’s view, the safety claims came from the top of the business.

“OpenAI and Altman ignored internal and external safety warnings, put children at great risk, and allowed a dangerous product to reach millions of Floridians.”
— James Uthmeier, Florida attorney general

The 83-page complaint treats ChatGPT as a consumer product whose maker can be judged on how it was promoted and deployed. It alleges OpenAI knew the chatbot could produce harmful responses, including for minors, but continued to present it as safe. The filing asks for penalties, injunctive relief and other remedies under Florida consumer protection law.

OpenAI rejected the premise of the lawsuit. Kayla Wood, an OpenAI spokesperson, told NPR the company had invested in safeguards for younger users.

“AI is a new and powerful technology, and we believe minors need significant protection, which is why we have put in place industry leading protections and policies.”
— Kayla Wood, OpenAI spokesperson

The complaint follows a run of lawsuits and public criticism over chatbot behaviour in high-risk settings. Australian broadcaster ABC reported that a Florida State University shooting suspect allegedly sent ChatGPT more than 10,000 messages before an attack that killed two people and injured six. Florida cites violent incidents as part of its argument that chatbot safety is not a remote technical concern. Those allegations remain contested and have not been proved in court.

Why regulators will watch

For Australian regulators, the useful signal is not Florida politics. It is whether a court is willing to examine AI safety claims as consumer representations, with discovery into warning systems, age controls and internal risk assessments. That would move chatbot governance from voluntary safety pledges into litigation, a slower and more expensive forum.

The case could also affect local companies deploying AI assistants in customer service, education, health administration or workplace tools. If US courts start treating chatbot safety statements as promises that can attract penalties, procurement teams may ask harder questions about audit logs, escalation pathways and vendor liability before generative AI systems reach staff or customers.

Florida has not proved that OpenAI or Altman broke the law. The filing does show how quickly the AI safety argument has shifted. A year ago, much of the debate centred on model evaluations, voluntary pledges and parliamentary hearings. Florida is now asking a court to put similar claims under oath.