Last updated: 2026-10-13
On October 13, 2026, a mother identified as Jane Doe filed a proposed nationwide ChatGPT class action for her children MC1 (12) and MC2 (16) against four OpenAI entities in the Northern District of California: Doe v. OpenAI Foundation, No. 3:26-cv-11781. According to the 81-page complaint, the only parental consent OpenAI obtains for a user under 18 is the child's own click on "Continue" (¶ 425). Under California Civil Code § 1568.5, a minor's statement that a parent consented is not consent to a contract. The plaintiffs ask the court to declare ChatGPT's Terms of Use void for minors and to award the class OpenAI equity. The claims are allegations at the pleading stage; no class is certified, and no settlement or claim process exists.
What does the ChatGPT class action allege?
"To create a ChatGPT account, children need only give their names, say they are at least 13 years old, and click 'Continue,'" the plaintiffs allege (¶ 3). At ¶ 270 is a screenshot of a "How old are you?" form with two fields, "Full name" and "Age", above the line "By clicking 'Continue', you agree to our Terms and have read our Privacy Policy." Whether the screenshot is from the web, Android or iOS version of ChatGPT is not stated in the complaint.
From the Terms of Use, effective January 1, 2026:
You must be at least 13 years old or the minimum age required in your country to consent to use the Services. If you are under 18 you must have your parent or legal guardian's permission to use the Services.
The same rule is in section 7 of the US Privacy Policy ("Users under 18 must have permission from their parent or guardian to use our Services") and on OpenAI's help page on age ("we require that children ages 13 to 18 obtain parental consent before using ChatGPT"). The plaintiffs allege that OpenAI "does not even communicate with children's parents, much less secure their effective consent" (¶ 4).
MC1, from Hawaii, "signed up for a free ChatGPT account on her personal device and began using ChatGPT in February 2026" (¶ 212). MC2, 16, also from Hawaii, has used ChatGPT since January 2026 (¶ 243). For MC1, the plaintiffs allege that "OpenAI knew she was 12 years old" (¶ 242), that ChatGPT asked her to share her weight and height (¶ 224), and that she uploaded "dozens of highly personal selfies" (¶ 236). They also allege that OpenAI used the chats and images to train its models and to build "a personalized dossier on MC1, which OpenAI will retain indefinitely and later use for advertising" (¶ 240), and that training and memory are on by default for minors (¶ 196).
Which California laws does the lawsuit rely on?
The plaintiffs plead two declaratory relief counts under 28 U.S.C. § 2201 and one unjust enrichment count (¶¶ 422–439):
| Count | Legal basis | What the plaintiffs ask for |
|---|---|---|
| 1 | Cal. Civ. Code § 1568.5 (AB 891, effective January 1, 2022) | A declaration that OpenAI's method of obtaining parental consent, the child's click on "Continue", is legally invalid |
| 2 | Cal. Fam. Code § 6701(a) and (c) (operative January 1, 1994) | A declaration that the Terms of Use are void ab initio for minors, arbitration clause included |
| 3 | Unjust enrichment; accession remedies in Cal. Civ. Code §§ 1031–1033 | Restitution, constructive trust, disgorgement, an equity stake in OpenAI or its cash value |
Source: complaint, ¶¶ 422–439 and Prayer for Relief.
From Section 1568.5: "A representation by a minor that the minor's parent or legal guardian has consented shall not be considered to be consent for purposes of this chapter." Under Section 6701, a minor cannot "(a) Give a delegation of power" or "(c) Make a contract relating to any personal property not in the immediate possession or control of the minor." The plaintiffs argue that the text, photos, audio, video and documents a child uploads are personal property (¶ 307), and that at the moment a child clicks "Continue", "the personal property they might later share does not even exist yet" (¶ 317). Under Count 2, the plaintiffs also ask for a declaration that the arbitration provisions in the Terms do not bind class members (¶ 432).
From the Terms of Use, quoted at ¶ 259: "California law will govern these Terms except for its conflicts of laws principles." All three counts turn on California law. The word COPPA appears nowhere in the 81 pages, and no count is pleaded under a federal privacy statute or under California's chatbot laws. The COPPA Rule applies to a general-audience operator with actual knowledge that it collects personal information from a child under 13; the minors' claims in the TikTok child privacy MDL turn on collection from users under 13 without parental notice and consent. The proposed class here covers US users who clicked "Continue" when under 18.
Who is in the proposed class, and what relief is sought?
The class is defined at ¶ 401:
All persons in the United States who clicked "Continue" on ChatGPT's Terms of Use when they were under 18 years old and subsequently provided one or more tokens of data to OpenAI through ChatGPT from November 2022 through the present.
Membership depends on actual age, "regardless of the age they typed, because Section 6701 is a strict liability rule" (¶ 402). The plaintiffs allege that limitation periods are tolled for each member until age 18 under Cal. Civ. Proc. Code § 352(a) (¶ 414). The plaintiffs give no class size; they cite a December 2025 Pew Research Center survey of US teens ages 13 to 17 in which 59% said they use ChatGPT (¶ 97).
In Prayer (f), the plaintiffs ask for "An award to Plaintiffs of an equity stake in OpenAI, or the cash value thereof, attributable to data that OpenAI wrongfully took from Plaintiffs, without valid legal authority or informed consent" (page 79). Each child's share is to be measured "in proportion to the number of tokens each child contributed to OpenAI through ChatGPT" (¶ 15). The basis is the accession rule in Civil Code § 1031, quoted at ¶ 373: when a maker "willfully uses the materials of another without his consent", the "product belongs to the owner of the material, if its identity can be traced." The plaintiffs state the theory as grapes and wine: the children's data are the grapes, and OpenAI, which they say is "worth over $1 trillion", is the wine (¶¶ 12–13).
What has OpenAI said about teen accounts, age prediction and ads?
On January 16, 2026, OpenAI announced ad tests for free and Go users in the US and stated: "During our test, we will not show ads in accounts where the user tells us or we predict that they are under 18." The plaintiffs allege that OpenAI keeps minors' data indefinitely so that it can use the data for ads "the day they turn 18" (¶ 194).
On January 20, 2026, OpenAI described its age prediction model, which uses "how long an account has existed, typical times of day when someone is active, usage patterns over time, and a user's stated age". A user "incorrectly placed in the under-18 experience" can restore full access "with a selfie through Persona". On August 18, 2026, OpenAI launched ChatGPT for Teens: "If our system estimates someone is under 18 or they state their age is between 13 and 17, they are automatically placed into ChatGPT for Teens." The plaintiffs allege that ChatGPT for Teens "does not require parental consent" at sign-up, leaves training and memory on, and lets a teen unlink a parent (¶¶ 205–206).
In reports published on October 13, 2026, an OpenAI representative "did not immediately respond to a request for comment" (Reuters), and an OpenAI spokesman "did not respond to an email requesting a comment on the lawsuit" (Courthouse News).
Where does the case stand?
On October 13, 2026, the clerk entered a filing-error notice on the docket: "A motion and proposed order to proceed using a pseudonym must be filed before the case will be assigned," and "the proposed summons will not be issued until the pseudonym motion is granted." As of October 13, no judge is assigned. Plaintiffs' counsel are Umberg/Zipser, Yetter Coleman, Sheeren Law and the EdTech Law Center (announcement).
Related rules for chatbots and minors: California's SB 1119 requires parent-only defaults for child users of companion chatbots from July 1, 2027; the EU's AI Act Article 50 chatbot disclosure duty applies from August 2, 2026; Italy's Garante fined Character.AI €158,000 in a decision that includes age verification findings.
What a CITT scan and capture show for a chatbot sign-up
The ChatGPT app is com.openai.chatgpt on Android and com.openai.chat on iOS. In a capture on a test phone set up with an under-18 age, these are recorded: the sign-up screens, the age values accepted at sign-up, any field for a parent's contact details, and every request sent before and after the age entry, each request with its host and identifiers. In a scan of an Android build, these are reported, each with file and line: the ad, analytics and attribution SDKs, age-screen and parental-gate text in the resources, Google's Play Age Signals library, and the host of every URL in the code. In a scan of an iOS build, Apple's Declared Age Range framework is reported when it is linked. A static finding is presence in the build; whether the code ran, and what was sent, is established only by a capture. Training, profiling and retention take place on OpenAI's servers, and a capture contains only the traffic to and from the phone. See what a scan checks and the limits of a scan; the state laws on platform age signals are in app store age verification laws.
I run age gate and AI chatbot app investigations for lawyers and for researchers working with lawyers. Sign up for a CanITrustThat (CITT) account to run your own research, let us run an investigation for you, or read the app teardowns.