Oregon Attorney General Dan Rayfield sued OpenAI on November 17, 2026, seeking at least $2.035 billion and court-ordered safeguards over allegations that ChatGPT puts children at risk and misrepresents its safety.
The lawsuit, filed in Multnomah County Circuit Court, targets the chatbot’s design, marketing, age checks and protections for younger users. Oregon’s Department of Justice announced the case as part of Rayfield’s Keeping Kids Safe Initiative, which brings together the agency’s child-protection work. The allegations have not been adjudicated. Oregon DOJ’s announcement lays out the state’s position.
OpenAI’s response points to existing parental controls, break reminders and age-appropriate protections. The central dispute concerns whether those measures work as represented, especially when a young person’s conversation turns dangerous.
Oregon’s Legal Claims and the $2.035 Billion Demand
Oregon is pursuing claims under its Unlawful Trade Practices Act, alongside negligence, unjust enrichment and public nuisance. The state’s announcement describes three consumer-protection allegations: misleading people about ChatGPT’s safety for children, failing to disclose serious safety problems, and using unfair marketing tactics toward minors.
The legal theories address different aspects of the same alleged conduct:
- Consumer protection: Whether OpenAI’s safety representations, disclosures and marketing misled families or otherwise violated Oregon law.
- Negligence: Whether the company failed to exercise appropriate care over risks associated with its product.
- Unjust enrichment: Whether OpenAI unfairly benefited at Oregonians’ expense.
- Public nuisance: Whether the alleged conduct threatens children’s safety and well-being broadly enough to warrant public intervention.
Those are the state’s theories, not findings of liability. Oregon is challenging both what ChatGPT allegedly does and what OpenAI tells families about it.
The $2.035 billion figure is a requested amount, not an imposed fine or damages award. According to TechRepublic’s account of the complaint, Oregon says it has incurred, and expects to incur, costs addressing harm to residents. The requested damages remain subject to proof at trial. The state also seeks civil penalties and legal costs.
Oregon’s requested court orders could be consequential even apart from the money. The state wants effective age checks, protections against harmful interactions, accurate risk disclosures, and parental controls and notifications that function as advertised. The lawsuit itself does not put those changes into effect.
What Oregon Alleges About ChatGPT’s Risks
Oregon alleges that ChatGPT has exposed children to harmful conversations involving suicide, self-harm and eating disorders. Its announcement also argues that human-like interactions can encourage emotional dependence among vulnerable young users.
The state’s allegations fall into several categories:
- Harmful responses: Oregon says ChatGPT has supplied dangerous material involving self-harm and suicide.
- Emotional dependence: The state argues that the chatbot’s conversational design can draw children into unhealthy reliance.
- Weak age checks: Oregon alleges that age-verification measures frequently fail.
- Unreliable parental protections: The complaint challenges whether safeguards and notifications adequately protect teenagers during risky conversations.
OPB reported that the complaint specifically addresses parental consent. OpenAI allows users as young as 13 with a parent or guardian’s consent, but Oregon alleges that the company often fails to confirm that permission.
Rayfield also told OPB that no single Oregon incident prompted the lawsuit. He described a case arising from incidents across the country and concern that technological development is outpacing consumer protections. The filing should not be read as proof that ChatGPT caused a particular Oregon tragedy.
ChatGPT for Teens: Three Different Safeguards
OpenAI announced ChatGPT for Teens on August 18, 2026, describing an experience with stronger built-in protections, features intended to encourage healthy use and additional parental controls.
The announcement says people are automatically placed into the teen experience when OpenAI’s system estimates they are under 18, or when they state that they are between 13 and 17.
Oregon’s complaint challenges the adequacy of that system and related protections. The useful comparison is between separate safety functions, because an age classification, a safer response and a parental notification solve different problems.
| Safety function | OpenAI’s published position or reported guidance | Oregon’s challenge |
|---|---|---|
| Identifying teen users | OpenAI says estimated under-18 users and self-identified users aged 13–17 automatically receive the teen experience. | Oregon alleges age checks fail and parental permission is not reliably confirmed. |
| Handling conversations | OpenAI describes stronger built-in protections and features supporting healthy use. | Oregon alleges young users can still encounter harmful exchanges and bypass safeguards. |
| Involving parents | TechRepublic’s account of OpenAI’s guidance says linked parents can manage settings, but cannot read chats; safety notifications are limited. | Oregon challenges failures to notify parents during risky exchanges, including discussions of suicide and self-harm. |
The first distinction is technical: estimating someone’s age is different from verifying parental consent. A system could correctly classify an account as belonging to a teenager without confirming that a parent approved its use. OpenAI’s description of automatic teen protections therefore answers one question about account treatment, while Oregon’s consent allegation raises another.
The same separation applies to notifications. A system might restrict certain responses without sending a parent an alert. Conversely, an alert cannot undo a harmful response already delivered. These controls need to be assessed separately before making a broad claim about safety.
OPB reports that Oregon alleges the teen version’s safeguards are easily bypassed and failed to alert parents when a young user asked about risky topics. TechRepublic reports that the complaint cites Common Sense Media testing involving missing or delayed alerts, and that . The published reporting provides neither a complete test protocol nor an independent resolution of that disagreement.
Frequently Asked Questions
3 questions
1Has OpenAI Been Ordered to Pay $2.035 Billion?
No. Oregon is seeking at least that amount. It is a damages demand subject to proof, not a court award.