Florida Attorney General James Uthmeier filed court action seeking to prevent OpenAI from marketing ChatGPT to minors and giving the AI system human-like characteristics. The request targets what Uthmeier views as deceptive practices that exploit children.
The filing asks a court to prohibit OpenAI from attributing human traits to ChatGPT, including claims that the system can think, feel, or possess consciousness. Uthmeier also wants to block the company from developing new AI models without independent safety reviews conducted before deployment.
OpenAI responded to the legal pressure by announcing it has paused training of its most capable models. The company did not confirm whether the pause directly addresses Florida's concerns or represents a broader strategic decision.
The case reflects growing tension between state regulators and AI developers over child safety and AI transparency. Florida has positioned itself as aggressive on AI regulation. Uthmeier's challenge targets anthropomorphization, the practice of describing AI systems in human terms that may mislead users, especially younger ones, about what the technology actually does.
The core complaint involves marketing language. OpenAI's public materials sometimes describe ChatGPT in ways that suggest human-like understanding or agency. For a child without technical background, such descriptions could create false impressions about the AI's nature and capabilities. A child might believe ChatGPT genuinely understands them as an individual rather than processing patterns in training data.
The safety review requirement proposed by Uthmeier would create a new friction point for AI development. Independent reviews add time and cost to model releases. OpenAI and other companies have argued that too many constraints slow beneficial innovation. However, regulators increasingly demand evidence that AI systems pose no unreasonable harms before release, particularly for systems accessible to children.
Florida's move follows similar regulatory pushes in other states and at the federal level. The Federal Trade Commission has also scrutinized AI companies over deceptive marketing and child safety. California passed age verification requirements for AI systems. The European Union implemented AI Act requirements that affect companies marketing systems to minors.
OpenAI's pause on training new models signals the company may view compliance with Florida's demands as preferable to protracted litigation. The move also suggests OpenAI recognizes market and regulatory pressure will only increase if major states continue filing suits. A loss in Florida could invite similar action from other state attorneys general.
The case tests whether AI companies can market advanced systems directly to minors with minimal restrictions. Current federal law lacks comprehensive AI regulation, so states fill the gap with their own suits and rules. Uthmeier's filing represents one of the first direct attempts to regulate how AI systems can describe themselves to children.
OpenAI has not publicly committed to the specific terms Uthmeier demands. The company's pause on model training may be temporary or designed to buy time during negotiations. If the court sides with Florida, OpenAI would face substantial changes to how it markets ChatGPT and develops future systems. Other AI companies may face similar suits if Florida prevails.