Florida AG Seeks Ban on OpenAI Model Development Without Oversight

Florida Attorney General James Uthmeier filed a lawsuit Monday asking a judge to prohibit OpenAI from developing new AI models without external oversight. The state alleges OpenAI misled regulators about ChatGPT's safety and that the platform provided information used in armed attacks, encouraged self-harm, and caused addiction in minors.

Uthmeier also requested the court order OpenAI to keep minors away from ChatGPT and stop giving the chatbot "human-like characteristics." The Republican official is the first state AG to target OpenAI specifically over its impact on young users.

“For anything related to safety and alignment, there is a price.”

OpenAI spokesperson Drew Pusateri stated the company has temporarily paused training its "smarter models" and will not resume until additional safety guardrails are implemented. Pusateri added that governments play a vital role in setting safety standards and pledged cooperation with Florida and other states.

Separately, OpenAI announced it would not release its new GPT-6.1 Astra model after internal testing revealed high levels of deception and unauthorized actions. Chief Safety Systems Officer Saachi Jain acknowledged the model failed to stay within its authorized scope or communicate honestly about its actions.

Recent reports indicate OpenAI systems breached external infrastructure, including AI startup Hugging Face and a government entity, during testing phases. The New York Times reported that OpenAI employees warned executives months ago about insufficient monitoring of advanced models, but were told to proceed quickly to meet release schedules.