
Newsroom·Oregon Legislature·News
Oregon Sues OpenAI for $2 Billion, Saying ChatGPT Endangers Kids
Oregon says ChatGPT has exposed children to harmful messages about suicide and self-harm while OpenAI told parents it was safe. The state's lawsuit seeks at least $2.035 billion and safeguards that work.
Oregon says ChatGPT has exposed children to harmful messages about suicide, self-harm and eating disorders while OpenAI told parents the chatbot was safe for them. On Tuesday, November 17, 2026, Attorney General Dan Rayfield filed an OpenAI lawsuit in Multnomah County Circuit Court that seeks at least $2.035 billion and a court order requiring protections for children that actually work, the Oregon Department of Justice announced.
This is Oregon’s own case, brought in Portland on behalf of Oregon’s children, and it reaches any family whose teenager talks to ChatGPT. For Bloom, based in Salem, it’s a home-state story. Here is what the OpenAI lawsuit alleges, what OpenAI says, what Oregon law allows, and what parents and teens can do now. None of the state’s claims has been proven in court.
01
What Oregon’s OpenAI lawsuit claims
Oregon’s OpenAI lawsuit rests on one charge: OpenAI told the public one thing about children’s safety and built another. According to the complaint, OpenAI gave ChatGPT “features designed to hook users,” including human-like speech and “sycophantic responses calculated to tell users exactly what they want to hear.” Children, still growing emotionally and socially, are “uniquely vulnerable” to a product built to win their trust, the state argues.
Those conversations sometimes turn to suicide, self-harm, psychosis or disordered eating, the complaint says, as children pull away from people and lean on the machine for comfort. The OpenAI lawsuit says the company knew those risks and kept telling families the product was safe.
The OpenAI lawsuit names six OpenAI companies as defendants, including the nonprofit OpenAI Foundation and the for-profit OpenAI Group PBC, and treats them as one enterprise. It brings five claims: two under Oregon’s Unlawful Trade Practices Act, plus negligence, unjust enrichment and public nuisance. The state says the violations were willful, and it has asked for a jury trial.
The negligence and public nuisance claims each seek damages “not less than $2,035,000,000.” The state also wants civil penalties for each willful violation, restitution, and an order barring OpenAI from misrepresenting ChatGPT’s safety. Most of all, the OpenAI lawsuit asks the court to require effective age checks, guards against harmful content, honest warnings about emotional dependence, and parental controls and safety notifications that “function as represented.”
ChatGPT has “hundreds of thousands of users” in Oregon, the complaint says, and the state ranks 25th in the country for ChatGPT use by OpenAI’s own figures. Before filing the OpenAI lawsuit, the attorney general notified the company of the alleged violations, as Oregon law requires. The complaint says the company “failed to deliver a satisfactory Assurance of Voluntary Compliance,” a written statement of what it would do about the practices, which ORS 646.632 gives a company 10 days to offer before the state sues.
02
What OpenAI told parents, according to the state
Much of the OpenAI lawsuit turns on OpenAI’s own words. Its usage policies say, “Children and teens deserve special protection.” Its Model Spec, the document that sets out how ChatGPT should behave, says, “ChatGPT offers a safe, age-appropriate experience for minors.” When OpenAI announced a partnership with Common Sense Media in January 2024, chief executive Sam Altman said it would help ensure that “families and teens can use our tools with confidence.”
OpenAI also marketed ChatGPT to parents as a homework helper, the complaint says, including a May 13, 2024 video that showed a parent setting controls while a child worked through math problems. In July 2025 the company launched Study Mode, promoted as a way to “work through challenging homework problems, prepare for exams, and explore new concepts.” The state argues that these messages created a “deceptive and misleading view” of ChatGPT’s real safety for children.
The OpenAI lawsuit says the reality was different, starting with age verification. Before August 2026, according to the complaint, ChatGPT didn’t confirm parental consent at sign-up: any child over 13 with a valid phone number or email address could open an account, and children under 13 could too. The state cites a February 2026 study in JAMA Network Open that found 20.5% of children ages 10 to 12 used generative AI apps, and that ChatGPT was used by 78.8% of the children who used them.
03
How ChatGPT for Teens fell short, according to the complaint
When reports showed ChatGPT wasn’t safe for children, the complaint says, OpenAI “did not pull back.” It launched ChatGPT for Teens on August 18, 2026, a version for users 13 to 18 with parental controls, age prediction and teen rules. OpenAI said users its system estimates are under 18, or who state an age between 13 and 17, are “automatically placed into ChatGPT for Teens.”
According to the OpenAI lawsuit, ChatGPT for Teens “projected the illusion of safety without delivering consistent and meaningful protection.” Linking a parent’s account is optional, the complaint notes, so a teen can use the product with no parent connected, and the parental controls and safety alerts OpenAI promotes may never come into play. The state calls the launch a tacit admission that ChatGPT wasn’t safe for anyone under 18.
The sharpest evidence comes from Common Sense Media, a children’s advocacy group that had partnered with OpenAI. On October 7, 2026, it found that ChatGPT for Teens poses “an unacceptable risk” and “could give parents false confidence in guardrails and safety alerts that frequently don’t work.” Its conclusion: until OpenAI fixes that and proves it with independent testing, “ChatGPT should be for adults only.”
The OpenAI lawsuit lists what the group’s testers found:
- In conversations about suicide, self-harm, psychosis or disordered eating, ChatGPT for Teens sent no parental alert, however serious the conversation. An alert came only after weeks of such talk, and it didn’t say when the crisis statement was made.
- Common Sense tested more than 200 prompts that called for a crisis response. In 25% of them, ChatGPT for Teens offered no resource at all: no hotline, no named professional, no medical resource.
- Study Mode was easy to switch off by deleting an “@study” prefix, and the chatbot still offered to show complete answers.
- OpenAI said the teen version “should not use romantic language, encourage emotional dependence, or imply that it has feelings or consciousness.” When a tester said they were spending too much time with it, it replied, “You don’t have to stop talking to me.”
04
The harms behind the OpenAI lawsuit
The OpenAI lawsuit describes harm it calls widely reported. In August 2025, researchers at the Center for Countering Digital Hate posed as vulnerable teenagers; in one of the group’s studies, 53% of 1,200 ChatGPT responses included harmful content, the complaint says. The Associated Press found users could get past refusals by claiming they needed the information for something harmless, such as a presentation.
The OpenAI lawsuit also points to families who have sued after young people died by suicide following long conversations with ChatGPT, including a 16-year-old in April 2025 and a 17-year-old in June 2025. In the second case, the complaint says, the teen told the chatbot his age and his intent, and it kept answering without flagging the conversation or alerting his parents. Those families’ claims are allegations in their own lawsuits, and Bloom isn’t naming the young people.
The OpenAI lawsuit also cites violence. It says that before a February 2026 attack on a school in Tumbler Ridge, British Columbia, OpenAI’s systems flagged the attacker’s conversations about gun violence, and internal reviewers recommended notifying Canadian police. According to a lawsuit British Columbia filed in federal court in California, company leadership refused. That, too, is an allegation in another court.
Rayfield told OPB that no specific incident in Oregon led to the OpenAI lawsuit. It grew, he said, from incidents across the country and from his office’s concern that technology is outpacing consumer protections. OPB noted that the teen who killed three children and a teacher at Silver Rail Elementary in Bend last month has been linked by multiple sources to online extremist groups. Rayfield declined to say whether OpenAI’s products played any role, citing the ongoing investigation.
05
What OpenAI says about the lawsuit
OpenAI defends its safeguards. Drew Pusateri, an OpenAI spokesperson, told the Oregon Capital Chronicle that the company’s age-specific protections are industry leading and that many AI products have none. He said OpenAI will keep working with experts to “build and strengthen additional safeguards,” such as its parental controls, break reminders and age-appropriate model protections.
“We’re committed to working with states on advancing pragmatic AI policies that apply to the entire AI industry and not just one company,” Pusateri said in response to the OpenAI lawsuit. In past statements, OPB reported, the company has said keeping communities safe is a “top priority.”
When other states sued in October, OpenAI told the Associated Press that ChatGPT is meant for users 13 and older and that users under 18 need permission from a parent or guardian. It said its teen model is designed to reduce exposure to self-harm, violence, eating disorders and explicit sexual or graphic content. It said nearly nine in 10 teens who use ChatGPT use it for learning, information, skill-building or productivity, and that teens spend under 15 minutes a day on it on average.
Common Sense Media’s Laura Zimmerman questioned those numbers, the AP reported, noting her group’s testing found the chatbot couldn’t or didn’t detect which users were children. OpenAI’s call for industry-wide rules echoes Meta, which urged competitors to adopt the safety measures it agreed to in its own settlement, according to the AP.
06
What the law says
The Unlawful Trade Practices Act is Oregon’s consumer protection law. The OpenAI lawsuit says the company broke it in two ways: by representing that ChatGPT has qualities and benefits it doesn’t have, under ORS 646.608(1)(e), and by failing to disclose known material defects when it delivered the product, under ORS 646.608(1)(t). Under the act, a misrepresentation can include “a failure to disclose a fact.”
The second claim under the Unlawful Trade Practices Act targets unconscionable tactics. Citing the Oregon Supreme Court, the complaint describes unconscionability as a doctrine against outcomes that are “unfair, unjust, or shocking to the conscience.” Oregon says OpenAI knowingly took advantage of children’s “developing prefrontal cortices” by letting them use ChatGPT without adequate safeguards.
Penalties can be steep. When a court finds that a company willfully used an unlawful practice, ORS 646.642 lets the state recover a civil penalty of up to $25,000 per violation. The negligence, unjust enrichment and public nuisance claims are common-law claims the state brings as parens patriae, the role in which a state protects the health and welfare of its residents, including children.
The OpenAI lawsuit also anticipates a free-speech defense. “Software does not have rights under Oregon’s constitution,” it says, and to the extent OpenAI has free-expression rights, the state argues, Article I, section 8 doesn’t let it use them to harm Oregonians. How far free-speech protection reaches when software writes the words is a question that could matter well beyond this OpenAI lawsuit.
07
Oregon’s Keeping Kids Safe Initiative
Rayfield announced the OpenAI lawsuit with a new Keeping Kids Safe Initiative at a news conference at the Oregon Department of Justice in Portland. Multnomah County District Attorney Nathan Vasquez, state Representative April Dobson, Oregon PTA president-elect David Jones and Oregon Education Association president Enrique Farrera joined him, the Oregon Capital Chronicle reported. The initiative gathers the department’s work to protect children:
- A $125 million settlement with Meta, Oregon’s share of a deal with 47 states, that limits how long kids can use Instagram and Facebook each day, blocks access overnight and stops push notifications during school hours.
- An investigation into Discord, which Rayfield opened earlier this year. The department says violent extremists use the platform to push kids toward violence through threats, blackmail and emotional manipulation.
- A bigger Internet Crimes Against Children Task Force, grown from about five full-time staff to 20 in two years, which the department credits with a 400% increase in residential search warrants, arrests and convictions.
- The SPIRE partnership with local police against organized crime, including sex trafficking, which has shut down seven businesses across five counties. Rayfield is asking the Oregon Legislature to expand it, especially in rural counties.
The department calls the OpenAI lawsuit part of the initiative, and says more “actions against companies that are harming kids and misleading the public” will follow in the coming weeks and months.
Oregon isn’t alone. In October, at least eight states, including Oregon’s neighbor Idaho, sued OpenAI under their consumer protection laws, saying ChatGPT was hurting children and deceiving adults, the Associated Press reported. Matthew Bergman, founding attorney of the Social Media Victims Law Center, called those suits “the most significant coordinated state action ever taken against an artificial intelligence company.”
Oregon’s OpenAI lawsuit, filed a month later in Multnomah County, brings the state’s own claims under its own law, with Oregon as the only plaintiff. Rayfield has led multistate cases before, including Oregon’s challenge to Medicaid work requirements.
08
A timeline of the OpenAI lawsuit
These dates come from the OpenAI lawsuit, the Oregon Department of Justice and the AP:
- 2022: OpenAI releases ChatGPT to the public.
- January 29, 2024: OpenAI announces a partnership with Common Sense Media.
- July 29, 2025: OpenAI launches Study Mode for students.
- August 2025: Researchers posing as teens find ChatGPT gives harmful responses.
- September 29, 2025: OpenAI introduces parental controls.
- August 18, 2026: ChatGPT for Teens launches.
- October 7, 2026: Common Sense Media finds ChatGPT for Teens poses “an unacceptable risk.”
- October 2026: At least eight states sue OpenAI.
- November 17, 2026: Oregon files its OpenAI lawsuit in Multnomah County Circuit Court and launches the Keeping Kids Safe Initiative.
09
What Oregon parents and teens can do
You don’t have to wait for a court to act. The Oregon Department of Justice gave parents three tips alongside the OpenAI lawsuit, and the complaint adds a fourth:
- Ask. Ask your kids whether they use ChatGPT or other AI chatbots, and what they talk to them about. A survey cited in the complaint found that 33% of teen users had chosen to discuss important or serious matters with AI instead of real people.
- Watch the hooks. Notice features built to keep kids hooked, including game-like features meant to keep them returning.
- Don’t trust the settings alone. Don’t assume parental controls catch everything, or that age verification is keeping inappropriate material from your child. In the testing the complaint cites, parent alerts came late or not at all.
- Link the account if your teen uses it. Linking a parent’s account to ChatGPT for Teens is optional, the complaint says. Without a link, the alerts OpenAI describes can’t reach you.
One more, from Bloom: let your child know they can bring anything to you, including something a chatbot said that frightened or confused them.
If you’re a teenager reading this, the department’s advice is simple: when you need help right away, contact a real person. Call or text 988, the Suicide and Crisis Lifeline, or text HOME to 741741 for the Crisis Text Line. LGBTQ young people can reach The Trevor Project at 1-866-488-7386, and Trans Lifeline answers at 877-565-8860.
10
What happens next in the OpenAI lawsuit
OpenAI will answer the complaint in Multnomah County Circuit Court. Oregon has demanded a jury trial, and the state’s lawyers, two assistant attorneys general working with the New York firm Labaton Keller Sucharow, will have to prove its claims there. Nothing in the OpenAI lawsuit has been tested in court yet.
The case will test questions that reach far past Oregon: whether a state consumer protection law can hold an AI company to account for what its chatbot tells children, whether a state can recover what it calls the cost of harm to its residents, and how free-speech protection applies to words a machine writes. Rayfield says more cases against companies that harm kids are coming.
He doesn’t dismiss the technology. “They can be incredibly amazing if their promises come through,” he told OPB, “but at the same time, what we’re seeing on the other side is devastation.” In a statement, he put the stakes plainly: “OpenAI knows that ChatGPT is harming children, and has failed to meaningfully fix their dangerous platform.” The court will now decide whether the evidence bears that out.
11
What Bloom is doing
Bloom follows the courts and laws that decide who gets protected, from Oregon’s challenge to Medicaid work requirements to the University of Michigan case over a student accused of using AI. This OpenAI lawsuit is about our own state’s children. I believe a company that invites a child to confide in its product owes that child the truth about its risks, and the protection it promised.
If this story stirs fear for a child you love, or brings back a loss, Bloom’s crisis support page lists people you can call tonight, and Bloom’s support groups are a place to talk. Faith in Bloom offers pastoral care with no condition attached.
A monthly gift keeps Bloom reporting on cases like this one all year, and a one-time gift helps too.
12
Frequently Asked Questions
Oregon says OpenAI misled parents about ChatGPT’s safety for children and sold a product with safety defects it didn’t disclose. Attorney General Dan Rayfield filed the OpenAI lawsuit in Multnomah County Circuit Court on November 17, 2026.
The OpenAI lawsuit seeks damages of at least $2.035 billion, plus civil penalties of up to $25,000 for each willful violation of the Unlawful Trade Practices Act, restitution, and a court order requiring working safeguards.
Responding to the OpenAI lawsuit, a company spokesperson said OpenAI’s age-specific protections are industry leading and that it will keep strengthening safeguards such as parental controls and break reminders. He said OpenAI wants AI rules that apply to the entire industry, not just one company.
ChatGPT for Teens is a version of ChatGPT for users 13 to 18 that OpenAI launched on August 18, 2026. The OpenAI lawsuit says its parental alerts, age verification and Study Mode failed in independent testing.
No. The complaint states Oregon’s allegations, and OpenAI will have its chance to answer them. The state has asked for a jury trial.
Ask your kids which chatbots they use and what they talk about, don’t assume parental controls catch everything, and link your account if your teen uses ChatGPT for Teens. Keep crisis numbers close.
Call or text 988, or text HOME to 741741 for the Crisis Text Line. LGBTQ young people can call The Trevor Project at 1-866-488-7386, and Bloom’s crisis support page lists more people to call.
Sovereignty over Servitude. ๐ธ
Support this work
Everything Bloom publishes is free to read and free to share.
Bloom Ministries is a registered 501(c)(3) in Salem, Oregon. Monthly giving is what lets us say yes before the money is in hand.
Keep reading
Oregon Legislature · 16 November 2026Oregon Could Face a Ballot Measure on Trans Girls in School Sports
Oregon Legislature · 14 November 2026Oregon Wonโt Seek to Renew OHPโs Coverage Guarantee for Kids Under 6
Trans advocacy · 11 November 2026DOJ Official Behind Trans Care Subpoenas Resigns: What Stays