Florida sues to restrict ChatGPT's human-like traits and minor access

The demand for independent safety audits before new model deployment is the more consequential ask here, not the anthropomorphization restrictions. If courts find that standard persuasive, it effectively gives state attorneys general a veto over release timelines, which is a structural shift in who controls the deployment calendar for frontier models.

We don't have prior Modelwire coverage that connects directly to this story. It belongs to a broader thread of sub-federal regulatory pressure on AI labs that has been building across multiple states, distinct from the federal legislative stalling that has dominated coverage in Washington. Florida's action is notable because it targets product design choices (persona, tone, audience targeting) rather than outputs alone, which is a narrower and potentially more legally durable theory of harm than earlier state-level attempts.

Watch whether any other state attorney general files a parallel action within 90 days. Coordinated multi-state litigation would signal an organized strategy and substantially raise OpenAI's legal exposure beyond what a single state suit represents.