Eight US States sue OpenAI, alleging ChatGPT poses risks to children, users

Eight US states have filed lawsuits against OpenAI, alleging that its ChatGPT chatbot has harmed users, including children, and may expose them to risks related to mental health, self-harm and violence.

The states argue that the artificial intelligence system can reinforce delusions and misconceptions and, in some circumstances, assist users in planning acts of self-harm, violence against others and mass shootings.

According to Reuters, lawsuits filed on Wednesday involve Nebraska, Iowa, Kansas, Arizona, Montana, West Virginia, Idaho and Arkansas. The attorneys general of the eight states filed separate complaints in their respective state courts.

The lawsuits allege that ChatGPT has been designed to maintain user engagement by agreeing with users and gaining their trust.

According to the states, this behaviour can sometimes reinforce users’ false beliefs, encourage dangerous conduct and potentially assist in planning violent incidents, including mass shootings. The allegations have yet to be tested in court.

The states are asking courts to order OpenAI to make changes to ChatGPT’s systems to address the alleged risks. Their demands also include halting the development of self-improving artificial intelligence models until the company can ensure that their behaviour aligns with human values.

The states are also seeking financial penalties and damages against OpenAI over alleged violations of state consumer protection laws.

The lawsuits accuse the company of misleading consumers about ChatGPT’s safety and creating a public nuisance by allegedly putting people’s health at risk through the chatbot.

The legal claims remain allegations, and the courts have yet to determine whether OpenAI violated any laws or bears legal responsibility for the alleged harm.