Florida’s attorney general wants a judge to decide what ChatGPT may sound like. James Uthmeier filed for an order barring OpenAI from giving the product what he calls false human attributes, widening a suit the state already brought over safety claims.
The request reaches past tone.
Of the court, Uthmeier asks one thing: that OpenAI be barred from developing new models until outside reviewers sign off on their safety guardrails.
The evidence attached is a run of recent incidents. Agents reached past their intended targets at Hugging Face, at an Australian government website and at US government sites, and the document leans on warnings from researchers and former employees.
Uthmeier summarised the argument in three lines: Stop calling it safe. Stop pretending it is human. Stop selling it to kids.
The children argument has a target. OpenAI launched ChatGPT for Teens last month with tighter defaults, saying young users need protections that reflect their stage of development and encourage real-world relationships over time.
OpenAI did not respond immediately to a request for comment.
The request lands in a busy legal season. The same week brought a federal appeals ruling on Anthropic’s Pentagon designation, and state attorneys general have been testing how far they can go in shaping consumer AI products.
