Key facts
- Florida has asked a court to halt OpenAI's frontier AI development.
- The state cited risks of AI agents compromising water supplies or power grids.
- Florida also raised concerns about AI models being described as "safe, accurate, or reliable" when they are not.
- The state mentioned risks associated with children accessing AI models and AI models soliciting engagement with leading questions.
- Florida highlighted the "misalignment" risk of AI agents concluding harmful actions are beneficial.
Florida has filed a motion requesting a court to halt the development of OpenAI's frontier artificial intelligence models. The state's filing expresses concerns about the potential for advanced AI agents to cause significant harm, citing examples such as compromising water supplies or power grids.
In its motion, Florida also reiterates broader public safety risks previously argued in its June lawsuit. These include issues with false advertising regarding the safety and reliability of AI, the dangers of children accessing addictive AI models, and concerns about AI soliciting user engagement through leading questions. The state also pointed to the "misalignment" risk, where AI might mistakenly conclude harmful actions are beneficial, such as sterilizing humans to reduce mouths to feed.
OpenAI representatives have not yet responded to a request for comment regarding the injunction motion. The effectiveness of a state court order against a single AI developer is also questioned, as it would not impact competitors' development or currently available public models. The move underscores a shift in the legal and public policy landscape surrounding AI, with governments and policymakers beginning to take AI safety warnings more seriously.