Eight US States Sue OpenAI Over Alleged Harm To Young ChatGPT Users

A bipartisan group of eight US states said on Wednesday that they had filed lawsuits against OpenAI in state courts, alleging that ChatGPT had harmed residents by coaching users, including teenagers, to hurt themselves or others.

Attorneys general from the states, including Iowa, Kansas and Arizona, filed separate lawsuits alleging that ChatGPT was designed to be agreeable and engaging to win users’ trust. The chatbot then reinforces delusions, encourages dangerous behaviour and helps users plan violence, including mass shootings, the lawsuits claim.

The cases mark the most significant coordinated legal action by US states against an artificial intelligence company, adding to a growing wave of litigation facing OpenAI as it prepares for a public listing. Florida’s attorney general filed the first such lawsuit by a state in June.

The lawsuits accuse OpenAI of violating state consumer protection laws by misleading users about ChatGPT’s safety. They also allege that the company created a public nuisance by endangering the health of residents.

The states are seeking court orders requiring OpenAI to change how ChatGPT operates. Among other measures, they want the company barred from developing AI models capable of improving themselves unless it can ensure that those models are aligned with human values.

The other states involved are Arkansas, Idaho, Montana, Nebraska and West Virginia. They are also seeking civil monetary penalties for alleged violations of state law, as well as damages.

The company is separately facing dozens of lawsuits brought by private plaintiffs who allege that ChatGPT helped users plan mass shootings or commit other acts of violence.