OpenAI Faces Class Action Over Children’s Data as Plaintiffs Seek Company Equity

A proposed nationwide class action filed in California federal court accuses OpenAI of training its models on children’s private ChatGPT conversations without valid parental consent. The remedy it seeks is barely conventional: a slice of the company itself, or its cash equivalent.

The 81-page complaint was lodged on October 13, 2026 in the U.S. District Court for the Northern District of California, case 3:26-cv-11781. A Hawaii mother is the named plaintiff, acting for her 12-year-old daughter and 16-year-old son and seeking to represent millions of other minors. No class has been certified and OpenAI has yet to answer in court; none of the allegations has been proven.

• Not a COPPA case: The plaintiffs bypass the federal children’s privacy statute, relying on California contract and property law instead.
• Unusual damages theory: They want minors to receive OpenAI equity proportional to the tokens each contributed.
• Part of a wider offensive: The filing lands the same week that eight state attorneys general sued OpenAI over alleged harms to young users.

A consent process reduced to one click

At the centre of the complaint is OpenAI’s sign-up flow: a child can create an account by entering a name, declaring they are at least 13 and pressing "Continue", even though the company’s own terms require parental permission under 18. OpenAI, the plaintiffs say, never contacts a parent to verify it exists.

The two siblings behind the case are identified only as MC1 and MC2. The girl began using ChatGPT in February 2026 at 12; her brother started in January at 16. She allegedly asked the chatbot about a pimple, shared selfies and asked it to identify her "ugly features"; he used it for homework. The complaint contends this is by design: ChatGPT validates emotions, mirrors a user’s writing style and leans on slang and emojis, while the memory feature rolled out in April 2024 lets it build on earlier disclosures. The proposed class covers everyone who accepted ChatGPT’s terms while under 18 and supplied at least one token of data since November 2022.

Why the lawyers avoided COPPA

COPPA shields children under 13 and is enforced by the Federal Trade Commission. This suit covers everyone under 18 and is brought by private litigants, so the attorneys turned to three California doctrines.

• Civil Code Section 1568.5, a one-sentence rule in force since 2022 holding that a minor’s own assertion of parental consent is not consent.
• Family Code Section 6701, traced in the complaint to 1874, bars minors from entering certain contracts. If the terms of service are void, so is the arbitration clause.
• Unjust enrichment, on the theory that OpenAI profited from something it had no right to take.

Grapes, wine and a claim on equity

The most striking element is the remedy. Alongside restitution and punitive damages, the plaintiffs want minors to receive a stake in OpenAI, or its cash value, proportional to the tokens each supplied. Their lawyers borrow an analogy from Roman law: steal grapes, make wine, and the victim is compensated for the wine, not the grapes. The complaint invokes California Civil Code Sections 1031 to 1033 on accession and puts OpenAI’s value above $1 trillion.

The sharper argument concerns design. Per the filing, OpenAI does not train on enterprise customers’ data by default and switches memory off for them, while for minors both are enabled from the start. The complaint cites Stanford researcher Jennifer King, who describes a "two-tier system" in which businesses enjoy stronger protection than the public.

The filing also targets ChatGPT for Teens, launched on August 18, 2026: it does not require parental consent at sign-up, leaves the terms unchanged and keeps training and memory on.

"If a stranger did this to a child, there would be outrage," said lead counsel Paul Yetter of Yetter Coleman, one of four firms behind the case.

OpenAI spokesperson Drew Pusateri has said the company built "industry-leading age-specific protections into our product that many other AI products lack", while acknowledging that no system is perfect.

The action is one front among many. Florida sued OpenAI and Sam Altman in June 2026 over ChatGPT’s safety, with minors at the centre, and eight states — Iowa, Kansas, Nebraska, West Virginia, Arkansas, Arizona, Idaho and Montana — have filed consumer-protection suits of their own. The Hawaii defendants include the OpenAI Foundation and OpenAI Group PBC, the entity created in the October 2025 restructuring.

The equity demand is a long shot as drafted, but the case need not win on that point to draw blood. Should a judge find the terms of service void as to minors — and the arbitration clause with them — claims would move into open court rather than a private forum. Certifying a nationwide class on the strength of one state’s laws remains the first hurdle. What the lawsuit exposes either way is an asymmetry that is hard to defend: if corporate data is valuable enough to be excluded from training by default, it is difficult to explain why a 12-year-old’s disclosures are not.