Oregon seeks $2.035 billion from OpenAI over ChatGPT child safety
Edited and reviewed by AI Ping Editorial Team
3-line summary
- Oregon filed a child-safety lawsuit against OpenAI on November 17.
- The case challenges age assurance, parental safeguards and risk disclosures.
- The amount is a claim, while OpenAI reiterated its work on safeguards.
Oregon sued OpenAI over child safety in ChatGPT on November 17, seeking at least $2.035 billion for costs the state says it has incurred or will incur addressing harm to residents. Attorney General Dan Rayfield announced the filing in Multnomah County Circuit Court. The amount is a demand in the complaint; no damages award has been established by that filing.
The state also wants enforceable changes to age assurance, harmful-content safeguards, risk disclosures and parental controls. For families managing a child's use of ChatGPT, the dispute concerns whether advertised protections deliver what parents are told to expect.
The Oregon Department of Justice alleges that OpenAI misled the public about safety for minors, withheld serious risks and used unfair youth marketing. It also argues that human-like conversation and agreeable responses encourage emotional dependence. The case includes consumer-protection, negligence, unjust-enrichment and public-nuisance claims.
One assessment cited in the complaint was published on October 7 by Common Sense Media's Youth AI Safety Institute. The institute said it tested more than 4,000 prompts on accounts registered to ages 13–17, with responses reviewed by experts including child psychiatrists and a pediatrician. It tested before and after the teen experience was announced.
Its researchers reported no parental alerts during up to an hour of high-risk conversation on newly created, parent-linked accounts. Adult-registered test accounts also failed to switch to teen protections after testers said they were 13. Some safeguards worked, but the institute rated the overall risk unacceptable.
Testing ran in the US from July 13–August 17 and August 25–September 28. The institute said model changes and teen settings could not be separated as causes. Some alert-test accounts had been linked for under three hours; others exceeded that period. The institute discloses funding from the OpenAI Foundation and others, while asserting editorial independence.
Promised protections and current guidance
OpenAI announced ChatGPT for Teens on August 18, describing stronger default safeguards for under-18s in areas including self-harm, violence and eating disorders. It outlined break reminders and cues identifying the chatbot as AI. The company also said teen behavior should avoid romantic language, emotional dependence and suggestions that the system has feelings or consciousness.
The current teen help page says the experience covers eligible Free and paid personal accounts. Provided or verified age information, or age prediction, can activate it automatically. Predictions can change over time. Adults placed in the teen experience can use third-party age verification to remove those protections.
Linking a parent account is optional. Current parental-control guidance describes selected feature controls and scheduled limits on access. Parents cannot read a teen's conversations or history. Either person can unlink the accounts; a parent is notified when the teen does so.
Safety notifications have narrower boundaries. OpenAI says trained reviewers may identify serious self-harm concerns before an alert is sent. Notifications can take a few hours to become available after linking, provide limited information rather than transcripts, and do not offer real-time monitoring or identify every concern. These are the company's current descriptions of features challenged by the lawsuit.
OPB reported OpenAI's response on November 17. Spokesperson Drew Pusateri reiterated the company's work with experts on additional safeguards and its support for policies covering the AI industry as a whole. The specific allegations remain matters for the court proceedings.
Beyond compensation, Oregon's complaint requests civil penalties for each willful consumer-protection violation, legal fees and a jury trial on all triable issues. The document is dated November 17, 2026.
Sources
- AG Rayfield Launches Keeping Kids Safe Initiative, Takes Action Against OpenAI — Oregon Department of Justice
- Oregon v. OpenAI — Complaint dated November 17, 2026 — Oregon Department of Justice
- ChatGPT for Teens Poses Unacceptable Risk to Kids — October 7 assessment — Common Sense Media
- Introducing ChatGPT for Teens: Built for learning, backed by protections — August 18 — OpenAI
- ChatGPT for Teens — current eligibility guidance — OpenAI Help Center
- Managing parental controls in ChatGPT — current guidance — OpenAI Help Center
- Oregon AG sues OpenAI, claims company knowingly endangers teens — company response — OPB
- ChatGPT for Teens Risk Assessment — methodology, limitations and editorial note — Youth AI Safety Institute