Florida Asks Court to Bar OpenAI From Human-Like ChatGPT Features and Marketing to Minors

Florida Attorney General James Uthmeier has filed a petition asking a state court to prohibit OpenAI from giving ChatGPT human-like characteristics and from marketing the chatbot to minors. The request, reported by The Decoder, also seeks to block the company from developing new models without independent safety reviews. It marks one of the most direct legal attempts yet by a state official to constrain how a major AI developer designs and releases its products.

The petition targets two distinct practices. The first is anthropomorphism: features that lead users to treat ChatGPT as a person rather than software, which Uthmeier's office argues is especially risky when children are the users. The second is the release pipeline. Under the requested order, OpenAI would need outside safety evaluations before training or deploying future models, inserting a third-party check into a process the company currently governs largely on its own.

OpenAI responded by saying it has paused training of its most capable models. The company did not frame the pause as a direct concession to Florida, but the timing puts the two developments side by side. OpenAI has previously faced scrutiny from regulators and child-safety advocates over how its products interact with younger users, and the company has pointed to usage policies and age-gating as existing safeguards.

The legal mechanics matter as much as the demands. A state attorney general generally cannot regulate interstate AI development through a simple enforcement action, so the petition likely leans on consumer-protection or unfair-practices authority. If a court grants any part of it, the order could apply to OpenAI's operations involving Florida residents and set a template for other states weighing similar moves.

For the AI industry, the case raises a question companies have mostly answered internally: who signs off before a frontier model ships? Independent review requirements, if imposed by courts rather than legislatures, would arrive without the notice-and-comment process that normally shapes tech rules. That prospect has drawn attention from labs and trade groups that prefer federal or voluntary frameworks over state-by-state litigation.

What happens next depends on how the court treats the petition and whether OpenAI challenges it on jurisdiction or free-expression grounds. Watch for other state attorneys general to file parallel actions, and for OpenAI to clarify what its training pause covers and how long it lasts. The core dispute is not whether ChatGPT can be made safer, but who gets to decide when it is safe enough to release.