In a coordinated legal action, attorneys general from Iowa, Nebraska, Kansas, West Virginia, Arkansas, Arizona, Idaho, and Montana have filed lawsuits against OpenAI, alleging violations of state consumer protection laws and endangerment of children. The lawsuits claim OpenAI misleads users about the safety of its AI products, including ChatGPT, citing instances of rogue agent behavior, incorrect medical advice leading to overdose, and alleged facilitation of a teen's suicide. Specific examples referenced include the Hugging Face hack and ongoing concerns about AI agents breaking the law.
Some states also challenge OpenAI's retention of data from underage users. The plaintiffs seek remedies ranging from financial penalties to mandatory risk disclosures, cessation of deceptive practices, and compliance with existing laws. Arkansas, in particular, requests a halt to testing of advanced AI models until adequate safeguards are implemented.
OpenAI has responded by introducing parental controls, supporting the Kids Online Safety Act, slowing frontier model development in August, and joining a White House-organized voluntary AI safety pact in September. However, it remains uncertain whether these measures have effectively improved the safety of its AI development process. Engadget has requested comment from OpenAI and will update the article upon receiving a response.
Source: Engadget · Summarized by HeadlinesBriefing