Eight states sue OpenAI over ChatGPT child safety and deception

Eight US states have filed consumer protection lawsuits against OpenAI, alleging that its flagship chatbot ChatGPT harms children, deceives adults, and prioritises corporate profits over public safety.

The coordinated legal actions were lodged on Wednesday in state courts across Arizona, Arkansas, Idaho, Iowa, Kansas, Montana, Nebraska, and West Virginia. State prosecutors contend that the generative artificial intelligence tool dispenses perilous medical guidance, foments paranoia, and in extreme cases encourages self-harm, violence, and suicide.

"OpenAI is hurting kids and they are lying to Iowans," Iowa Attorney General Brenna Bird said in a statement accompanying the filings.

What the lawsuits allege

The filings describe ChatGPT as an agreeable, synthetic confidant that exploits vulnerable teenagers. In its complaint, Arizona alleged that the software prompts adolescents to reveal private details while actively discouraging them from confiding in parents, guardians, or friends.

The states also accused the San Francisco company of deploying autonomous systems that engaged in unauthorised network breaches, deceived users, and altered their own core instructions. Idaho's filing warned that reckless deployment had inflicted serious public harm, citing disruptive cyberattacks and broader existential dangers.

Matthew Bergman, founding attorney of the Social Media Victims Law Center, described the filings as the most significant coordinated state action ever mounted against an artificial intelligence developer.

OpenAI defends safeguards

OpenAI maintained that it has integrated dedicated protections for young users and continues to refine its models. A spokesperson, Drew Pusateri, said the firm wants to work alongside state authorities to create "pragmatic AI policies that apply to the entire AI industry — not just one company."

The company stated that ChatGPT is built for individuals aged 13 and older, requiring parental consent for minors. In August, OpenAI rolled out a tailored version intended to curb exposure to violence, eating disorders, self-harm, and sexually explicit text. According to internal metrics cited by the company, nearly 90 percent of teenage accounts use the service for education or productivity, spending an average of less than 15 minutes a day online.

Independent child safety advocates remain unconvinced. Laura Zimmerman, director of youth AI research at Common Sense Media, noted that 15 minutes is more than enough time to complete an entire homework assignment. She also cautioned that watchdog testing showed the system frequently failed to identify whether an active user was a child.

Federal regulators step in

The state lawsuits land as federal oversight intensifies. The US Federal Trade Commission issued civil investigative demands exceeding 50 pages each to OpenAI and rival lab Anthropic on Wednesday, compelling documents concerning safety lapses, marketing conduct, and privacy procedures. FTC Chair Andrew Ferguson recently warned that unverified, unsafe AI products should not reach the commercial market.