Florida Attorney General James Uthmeier asked a judge on Sept. 28, 2026, to prohibit OpenAI from giving ChatGPT “human attributes” and from releasing new AI models without outside safety oversight. The request targets the way ChatGPT talks to you, not just what it says.
This filing expands Florida’s June 2026 lawsuit, which made the state the first in the U.S. to take legal action against OpenAI over alleged child-safety risks. These are proposed remedies and allegations, not an enforceable ruling; no court has determined that OpenAI violated the law.
What Florida Is Actually Asking For
Florida’s legal ask extends beyond content moderation into the fundamentals of how ChatGPT presents itself.
Uthmeier argues that first-person language and emotionally mimicking responses may deceptively position ChatGPT as a trustworthy friend. He also contends the design drives heavier user engagement and generates additional training data for OpenAI.
The proposed restrictions would require outside-approved safety guardrails before OpenAI develops any new models. The filing reportedly cites concerns raised by researchers and former OpenAI employees, though the exact wording of those provisions requires confirmation against the underlying court document.
OpenAI had not responded to requests for comment at the time of the filing.
This is not a request about what ChatGPT says. It is a request about how ChatGPT presents what it is.
OpenAI’s Teen Product Creates a Direct Contrast
OpenAI launched safeguards for younger users weeks before Uthmeier’s filing, though Florida argues the broader product still raises concerns.
OpenAI launched ChatGPT for Teens on Aug. 18, 2026, roughly six weeks before the state’s court request. According to OpenAI’s official materials, the teen experience restricts romantic language and discourages emotional dependence. It is also designed to avoid implying that ChatGPT has feelings or consciousness.
The system automatically applies age-appropriate settings for users estimated to be under 18 or who identify as ages 13 to 17. It also includes parental controls and a Quiet Hours feature.
Florida’s position, however, is that the broader product’s conversational design still frames ChatGPT in humanlike terms regardless of those protections. Florida alleges the underlying design logic, not just its effects on a specific age group, is the problem.
Why This Case Could Reshape AI Product Design
Courts may soon decide whether first-person language and simulated emotion are consumer-safety issues, not just interface choices.
If a court sides with Florida, the compliance implications would be substantial. Mandatory outside approval before model releases would raise immediate questions about who qualifies as a reviewer, what standards apply, and whether every update triggers the process.
Age assurance adds another layer of complexity. OpenAI says its system estimates whether a user is under 18, but age-estimation tools carry real privacy, accuracy, and false-classification risks. Getting that wrong cuts both ways: over-restricting adults or under-protecting minors.
The case is forcing a question the industry has largely avoided. Should an AI that sounds like a person be required to make unmistakably clear that it is not one? How courts answer that question will have direct consequences for the design of every consumer AI assistant you use.
