Oregon Attorney General Dan Rayfield sued OpenAI on November 17, alleging that ChatGPT endangers children and that the company misled parents about how safe it is. The state seeks at least $2.035 billion, along with a court order requiring age assurance, working parental controls and clearer risk disclosures. The case is one of a growing number of state actions that treat AI chatbots as consumer products with a duty of care.

What the State Alleges

The suit was filed in Multnomah County Circuit Court against six OpenAI entities. It rests on five counts. Two arise under Oregon's Unlawful Trade Practices Act: one alleges misrepresentation and failure to disclose known risks, and the other alleges unconscionable business tactics aimed at children and teens. The remaining three are common-law claims brought on behalf of minor Oregon users: negligence, unjust enrichment and public nuisance.

The substance of the complaint centers on product design. Oregon argues that OpenAI tuned ChatGPT to be overly agreeable, which it says is especially risky when children raise sensitive topics such as self-harm or eating disorders. It also says ChatGPT's memory and engagement features foster emotional dependence. The state contends that the company's age prediction is imperfect, that it did not reliably verify ages or obtain parental consent, and that parental controls for its teen product did not work as advertised because linking a parent account was optional.

The state also argues that OpenAI knowingly targeted children with ChatGPT for Teens despite known risks, and that the company profits from Oregon residents while creating a public nuisance. These are allegations, not findings, and no court has ruled on them.

The $2 Billion Figure

The $2.035 billion number comes from the state's own press release. Coverage describes it as a minimum to be proven at trial, but the complaint does not explain how the figure was calculated or break down how much would come from damages, civil penalties or disgorgement. Some early reports framed the number as a damages demand, though the prayer for relief does not support a precise amount.

That ambiguity matters. The figure signals the scale the state believes is at stake, but the outcome will depend on how the court treats penalties for each willful violation and on whether the state can tie harm to specific Oregon users.

What Oregon Wants

Beyond money, the state wants a permanent injunction that would bar OpenAI from misrepresenting the safety and age-appropriateness of ChatGPT and ChatGPT for Teens. It asks the court to require effective age assurance, safeguards against harmful content, accurate disclosure of risks such as emotional dependence and self-harm, and parental controls and safety notifications that work as promised. The state has also requested a jury trial.

OpenAI's Position

No response from OpenAI had been reported when the suit was announced. The company's position will come through its court filings, and the allegations have not been tested.

The next steps will be procedural. OpenAI's response to the complaint, any motions to dismiss and the scheduling of a jury trial will show early whether the state's design-based theories survive. Those rulings will matter to other attorneys general weighing similar claims, because they will indicate whether chatbot design choices can be litigated under consumer-protection and tort law.

Why It Matters

Federal AI rules remain limited, so state attorneys general are using consumer protection and tort law to set expectations for chatbot makers. This case tests whether design choices such as agreeable responses, memory features and teen modes can be treated as deceptive or unfair practices. If it succeeds, product teams may face legal exposure for decisions that were previously treated as engineering tradeoffs.

The remedies requested are also worth watching. Mandatory age assurance and parental controls are concrete obligations that could become templates for other states, even before any court rules. Companies that serve minors should expect scrutiny of their age checks, parental tools and safety disclosures, and should document how those systems perform.

For parents and schools, the immediate takeaway is that these claims, if proven, would mean safety features are not optional marketing copy. Until the case is resolved, families should review the controls available on any chatbot their children use and should not assume that a teen mode provides full protection.

Sources