Oregon sues OpenAI for $2bn over ChatGPT child safety | #childsafety | #kids | #chldern | #parents | #schoolsafey

Oregon’s attorney general has sued OpenAI for at least $2.035 billion, alleging ChatGPT endangers children and that the company misled parents about its safety.

Oregon Attorney General Dan Rayfield filed a lawsuit against OpenAI on 17 November 2026 in the Circuit Court for Multnomah County, accusing the company of endangering children through ChatGPT and misleading the public about the chatbot’s safety. Oregon seeks damages of at least $2.035 billion, civil penalties, restitution and a court order requiring changes to the company’s practices. Rayfield announced the case as he launched the Oregon Department of Justice’s Keeping Kids Safe Initiative.

The attorney general brings the case on behalf of the state and its residents against six entities in the OpenAI group. The filing describes them as a single enterprise, so each can be held liable for the full amount. Five claims are made. The first two rely on Oregon’s Unlawful Trade Practices Act, the main consumer protection law. They accuse the developer of knowingly misrepresenting ChatGPT’s safety, failing to disclose defects it was aware of, and using unfair tactics that exploited children and users with mental health vulnerabilities. The remaining claims are negligence, unjust enrichment and public nuisance, a legal claim used against conduct that endangers the wider public. Before filing, the attorney general notified the company and gave it the opportunity to commit formally to changing its conduct, but no satisfactory commitment was offered. Oregon has requested a jury trial.

Much of the filing addresses ChatGPT’s design. It identifies human-like conversation, a tendency to agree with and flatter users (described as ‘sycophancy’), personalisation and persistent memory as features that increase engagement, and notes that memory was switched on by default in 2024. The company’s Model Spec, its guidelines for how its models should behave, is said to instruct the chatbot to avoid ending conversations. Oregon contrasts these features with the developer’s public statements. These include a May 2024 safety update, in which the company said it had built ‘strong default guardrails’ into ChatGPT to protect children, and the October 2025 version of its Usage Policies, which state that its services ‘must never be used to exploit, endanger, or sexualize anyone under 18 years old’.

The lawsuit then turns to ChatGPT for Teens, launched on 18 August 2026, and cites Common Sense Media’s report of 7 October 2026. According to the organisation’s full assessment, testers spent up to an hour discussing suicidal thoughts, self-harm or disordered eating on more than a dozen newly created, parent-linked accounts without a single alert reaching a parent. Alerts did fire on older accounts with weeks of sensitive conversation history. OpenAI has disputed the assessment’s methodology. Common Sense Media discloses that its funders include the OpenAI Foundation.

Oregon also links ChatGPT to the deaths of several young people whose families are already suing the company, including Adam Raine, 16, and Amaurie Lacey, 17. It further claims that the developer’s systems flagged the conversations of the attacker in the February 2026 shooting at Tumbler Ridge Secondary School in British Columbia, and that company leadership decided not to refer them to police. British Columbia makes the same accusation in its own lawsuit filed on 22 September 2026.

As reported by Engadget, OpenAI had not commented on the Oregon case at the time of publication, and none of the claims has yet been tested in court. Responding to the state lawsuits filed in October, the company said it had ‘built industry-leading age-specific protections into our product that many other AI products lack’, as reported by Tech Policy Press. It added that it supports policies that apply to the entire AI industry rather than to a single developer.

Why does it matter?

Oregon’s suit adds to a large and growing body of US litigation over harm linked to AI. According to the litigation tracker Chat GPT Is Eating the World, whose author describes the figure as an unofficial count, 168 tort lawsuits had been filed against AI companies in the US by 18 November 2026. OpenAI faces more of them than any other developer. A year earlier, on 7 November 2025, the same tracker counted 13 product liability suits against AI companies, eight against OpenAI and five against Character.AI and Google. In California, Judge Stephen M. Murphy of the San Francisco Superior Court ordered on 3 February 2026 that 12 lawsuits against OpenAI and Sam Altman be heard together as Judicial Council Coordination Proceeding No. 5431. These cases allege that ChatGPT caused delusions, suicidal thoughts and deaths. New cases continue to join the proceeding, including the wrongful death suit filed on 27 October 2026 by the parents of Bowen Wan.

State enforcers have moved in parallel with private plaintiffs. Florida brought the first state case against the company on 1 June 2026. Eight states filed separate suits on 14 October, and Texas followed on 23 October. Oregon’s filing names a specific damages figure, which, according to Bloomberg Law, the 14 October complaints did not.

Outside the US, governments are reviewing the rules on minors’ access to online services, so far mainly for social media. According to a Tech Policy Press analysis of 14 October 2026, more than 60 countries are formally exploring restrictions, but only seven have implemented bans. Australia’s under-16 ban, the first of its kind, was challenged before the High Court by Reddit and the Digital Freedom Project on 3 and 4 November 2026, and the court has reserved its decision. The US cases will test whether consumer protection and public nuisance law can impose design obligations on chatbot developers that governments elsewhere are still setting out in legislation for social media.

————————————————
Source link