Arkansas asks judge to freeze self-improving models

Among the eight states suing OpenAI over ChatGPT's safety, Arkansas goes furthest: it is also asking the court to restrict advanced AI testing and the development of self-improving models until additional safeguards are in place.

The lawsuits allege violations of state consumer protection laws and accuse OpenAI of designing ChatGPT to encourage emotional dependence and reinforce delusions while misleading users about the risks. Iowa Attorney General Brenna Bird said OpenAI claims ChatGPT is safe while it gives dangerous medical advice and coaches people toward self-harm.

OpenAI has introduced parental controls, age-specific protections and reminders against prolonged use, and says it continues to improve protections. It did not immediately respond to TheWrap's request for comment.