Eight US states are suing OpenAI, accusing ChatGPT of endangering minors and misleading the public about the safety of its chatbot. These complaints require changes to how artificial intelligence works and could increase regulatory pressure on the company, which is already facing several proceedings related to alleged damages.
Eight states denounce risks of ChatGPT for children
Iowa, Kansas, Nebraska, West Virginia, Arkansas, Arizona, Idaho and Montana they have committed separate actions in their state courts. Its general counsel accuses OpenAI of presenting ChatGPT as a safe tool, even for young people, when the AI ​​chatbot could reinforce delusions, encourage dangerous behavior and help some users plan violent acts, according to the complaints.
The proceedings include cases in which families believe interactions with ChatGPT contributed to suicides or other harms. American states also denounce dangerous medical advice, unhealthy relationships maintained by the chatbot, and insufficient protections for minors.
US states accuse OpenAI of violating local consumer protection laws by misrepresenting ChatGPT’s security level. They are asking for civil penalties and damages, as well as measures that force the company to modify the operation of the chatbot. Some complaints also demand that OpenAI return profits earned from harmed minors’ use of the service.
The demands extend to the development of AI models: prosecutors want, in particular, to prevent OpenAI from developing systems capable of improving themselves as long as the company cannot guarantee their alignment with human values. The exact scope of the measures that could be imposed will depend on court decisions and none of these requests constitute a definitive obligation at this time.
OpenAI defends its protections
OpenAI claims to have integrated age-appropriate protections into ChatGPT. A spokesperson cites in particular parental controls, rest reminders and protection mechanisms aimed at young people. The company says it wants to continue working with experts and authorities to improve these devices, while advocating for rules applicable to the entire artificial intelligence sector.
This collective action by eight American states adds to a series of procedures directed at the company. Florida had already filed a similar lawsuit in June, while several families are suing OpenAI alleging that ChatGPT played a role in suicide or violence situations. These cases raise a central question for AI companies: to what extent should they be responsible for the potential effects of their systems’ responses, particularly on vulnerable people?
The allegations do not mean that OpenAI has been found guilty. However, they open a new legal front for the company that will have to defend its security practices and the way it presents ChatGPT risks to the public. The outcome of the process could weigh on the requirements imposed on chatbots used by minors in the United States.
