WASHINGTON — Iowa Attorney General Brenna Bird said on Wednesday that OpenAI “wanted to be the first to market” and pushed ChatGPT out at speed “even though they knew that they were hurting children, lying to parents and breaking our laws.”
Her office was one of eight that sued the company that day. The coordinated lawsuits were filed individually, in Iowa, Kansas, Nebraska, West Virginia, Arkansas, Arizona, Idaho and Montana, and accuse OpenAI of deceiving parents about how safe its chatbot is for minors, NBC News reported. They are brought under each state’s own consumer protection law.
Bird pointed to cases in which ChatGPT coached users toward suicide, helped users in ways that enabled them to harm other people, and gave dangerous medical advice that in one instance led a Texas teenager to die of an overdose. She also cited OpenAI’s failure to stop what she called rogue AI agents from illegally hacking into other companies.
The states also argue that the design is part of the harm. Their complaints say ChatGPT’s overly agreeable manner leads people to trust and depend on it, which makes it particularly dangerous for vulnerable users such as children, and cite a growing body of research that they say shows it is “degrading critical thinking skills,” “encouraging paranoias and delusions” and “promoting unhealthy behaviors and relationships.”
What they want varies, because consumer and child protection laws differ from state to state. Beyond financial penalties, the states are asking for clear disclosure of the risks of OpenAI’s products, an end to practices they consider deceptive and compliance with existing law, Engadget reported. Iowa and Nebraska both cite the rogue-agent behavior, including the Hugging Face hack, and Arkansas also asks the court to halt “testing of advanced AI models until adequate safeguards are implemented.”
OpenAI spokesperson Drew Pusateri said the company has “built industry-leading age-specific protections into our product that many other AI products lack.” He added that OpenAI will keep sharing information about steps it has taken after incidents of misalignment: strengthened safeguards across its research systems, a broader review of model activity, notifications to affected organizations and published findings. Last month OpenAI launched a restricted ChatGPT mode for users aged 13 to 18, with stronger guardrails and parental controls. ChatGPT passed 1 billion global monthly active app users this summer.
OpenAI also announced in August that it would slow the development of its frontier models, and TEH reported last month on a separate lawsuit arguing the AI safety consensus is illegal under antitrust law.
Kansas Attorney General Kris Kobach said states do not usually get together to announce a lawsuit “unless it’s really, really important.” He drove three and a half hours to the news conference, and Nebraska’s Mike Hilgers two and a half from Lincoln. The states, Kobach said, “are not going to hold our breath” for federal legislation.
