Published
July 1, 2026
A proposed class action accused OpenAI of quietly routing ChatGPT users' queries to Meta and Google through embedded tracking code. The named plaintiff dropped the case eight days after filing it.
Allegations Only · Case Voluntarily Dismissed
This article describes a class action complaint. The statements below are unproven allegations from that filing. OpenAI has not been found liable, no class was ever certified, and the named plaintiff voluntarily dismissed the case shortly after filing it. There is nothing to claim. This page is informational and is not legal advice.
What Is This About?
On May 5, 2026, a California resident named Saje Lim filed a proposed class action against OpenAI Global, LLC in the U.S. District Court for the Northern District of California. The case, Lim v. OpenAI Global, LLC, No. 3:26-cv-04063, accused OpenAI of secretly sharing ChatGPT users' private queries and identifying information with Meta and Google. Lim's attorneys, the consumer-litigation firm Bursor & Fisher, P.A., asked the court to certify a nationwide class of ChatGPT users and a California subclass.The case did not stay open for long. Court records show the plaintiff voluntarily dismissed it without prejudice on May 13, 2026 — eight days after filing — and the court closed the case two days later, on May 15, 2026. Nothing in the docket explains why. A voluntary dismissal without prejudice is not a ruling on the merits: it does not mean a judge rejected the claims, and it does not mean OpenAI was cleared of anything. It simply means the plaintiff withdrew the case, and the same allegations could, in theory, be refiled later — by Lim or by a different plaintiff represented by different counsel.
Status
Voluntarily Dismissed — Case Closed
Filed May 5, 2026 · dismissed without prejudice May 13, 2026 · terminated May 15, 2026
What Was Alleged
Meta and Google tracking code on ChatGPT.com
Meta Pixel and Google Analytics allegedly forwarded query content and identifying cookies
Can I Claim?
No — no class was certified and no settlement exists
What the Complaint Alleged OpenAI Did
The complaint's theory rested on two pieces of widely used advertising technology: Meta's "Meta Pixel" and Google Analytics. Both are snippets of code that a website operator can add to its own site so that, according to the complaint, a visitor's browser quietly sends a second, parallel message to Meta's or Google's servers every time it loads a page — separate from, but triggered by, the same request that loads the page for the user. The complaint alleged OpenAI added both to ChatGPT.com.According to the complaint, that meant two things happened whenever someone typed a question into ChatGPT: OpenAI's own servers processed the query to generate an answer, and — the complaint alleged — a parallel copy of information about that activity went to Meta and Google at the same time, without a pop-up, banner, or separate consent screen calling it out.
What Information the Complaint Says Was Shared
The complaint pointed to specific data points it said were transmitted alongside each other:• To Meta, the complaint alleged OpenAI's Meta Pixel code sent a version of the query topic (using the page title generated in the user's browser tab as its example) together with browser cookies — including a "c_user" cookie the complaint said carries a user's unencrypted Facebook ID — that Meta could use to look up a specific Facebook profile.
• To Google, the complaint alleged OpenAI's Google Analytics code sent query information together with a hashed version of a user's email address and Google account identifier cookies, which the complaint said Google — as the party that generates the hash — can match back to an existing Google advertising profile.
The complaint argued that because both companies already maintain detailed advertising profiles on billions of users, tying ChatGPT activity to a Facebook ID or a Google account effectively identified the person behind an otherwise anonymous-looking query. It cited Meta's and Google's own public documentation describing how the Meta Pixel and Google Analytics collect and use this kind of data for ad targeting, along with outside commentary questioning whether "hashed" identifiers are truly anonymous.
The Legal Claims
The complaint asserted four counts:• Electronic Communications Privacy Act (18 U.S.C. § 2511) — the federal wiretap statute, alleging OpenAI intentionally intercepted and disclosed users' electronic communications to Meta and Google without authorization.
• California Invasion of Privacy Act, Section 631 — alleging OpenAI aided, agreed with, or conspired with Meta and Google to read the contents of users' communications while they were in transit, without the consent of all parties.
• California Invasion of Privacy Act, Section 632 — alleging OpenAI used an "electronic recording device" (the tracking code) to eavesdrop on communications California subclass members reasonably expected to remain confidential.
• Invasion of privacy under the California Constitution and common law — alleging the alleged conduct amounted to an intrusion upon class members' reasonable expectation of privacy.
Setting the Record Straight: What the Complaint Actually Asked For
Because AI-generated summaries of this complaint have circulated online, it's worth being precise about what the filing itself requested, rather than what a paraphrase implies. The complaint's prayer for relief asked the court for six things: class certification under Rule 23; an injunction barring OpenAI from continuing the challenged conduct; a finding of liability on all four counts; damages (actual, statutory, punitive, and nominal, in an amount to be determined at trial); attorneys' fees and costs; and prejudgment interest. The complaint's injunction request is narrower than some summaries suggest — it asks the court to stop the alleged conduct going forward, but the filing itself does not include an explicit demand that OpenAI delete or destroy previously collected data.On damages, the complaint sought statutory damages of $10,000 or $100 per day per violation under the federal wiretap statute, and — under the California Invasion of Privacy Act — the greater of $5,000 per violation or three times actual damages. Those are the statutory caps set by the two laws the complaint cited; they are not amounts a court awarded, since the case never reached a ruling on damages.
On the proposed class itself, the complaint's opening "Nature of the Action" section described the case as being brought on behalf of "all United States residents" who used ChatGPT.com. The complaint's later, more formal Rule 23 class-definition paragraph is written slightly more broadly — "all persons who ... had their personally identifiable information and communications with ChatGPT disclosed to third party entities" — without repeating the U.S.-residency phrase. Both descriptions point to the same practical group of ChatGPT users; the point is only that the complaint itself is not perfectly consistent about the exact boundary, and no court ever had to resolve the difference because no class was certified.
Free settlement alerts
Get notified when new class actions open to claims
Join thousands of readers who get the latest class action settlements you may qualify for — delivered straight to your inbox.
For more class actions keep scrolling below.
Why This Matters Even Though the Case Closed
A dismissed complaint is not proof of anything, in either direction. But the underlying legal theory — that embedding third-party ad-tracking code like the Meta Pixel or Google Analytics on a site that handles sensitive conversations can trigger wiretap liability under CIPA and the ECPA — is not unique to this case. OpenClassActions.com is tracking a closely related theory currently being litigated against a different AI chatbot in the Perplexity AI chat privacy class action, and the same statutes have already produced class settlements against other companies, including the GameSpot CIPA settlement. Readers who want background on the underlying California statute can see OCA's CIPA glossary explainer. This case is also one of several unrelated OpenAI lawsuits making news this year — see OCA's roundup of which OpenAI lawsuits are actually class actions for how it compares to The New York Times' copyright suit and the consolidated author copyright cases.Frequently Asked Questions
What was Lim v. OpenAI Global, LLC about?
A proposed class action filed May 5, 2026 in the Northern District of California alleged OpenAI embedded Meta Pixel and Google Analytics tracking code into ChatGPT.com, which allowed Meta and Google to receive users' query content along with identifying cookies, without users' knowledge or consent.
Is the OpenAI ChatGPT privacy lawsuit still active?
No. The named plaintiff voluntarily dismissed the case without prejudice on May 13, 2026, eight days after filing, and the court terminated the case on May 15, 2026. A voluntary dismissal without prejudice does not decide the allegations on their merits and does not prevent the same claims from being refiled.
What information did the complaint say OpenAI shared with Meta and Google?
The complaint alleged OpenAI's Meta Pixel code transmitted a user's query topic alongside Facebook identifier cookies, and that OpenAI's Google Analytics code transmitted query topics alongside a hashed email address and Google account cookies, allowing Meta and Google to link ChatGPT activity to a user's existing advertising profile.
What laws did the complaint allege OpenAI violated?
The complaint asserted four claims: the federal Electronic Communications Privacy Act (18 U.S.C. § 2511), California Invasion of Privacy Act Section 631, California Invasion of Privacy Act Section 632, and invasion of privacy/intrusion upon seclusion under the California Constitution and common law.
Can I file a claim in this lawsuit?
No. The case was voluntarily dismissed before any class was certified and before any settlement was reached, so there is no claim form and nothing to file. This page will be updated if the same allegations are refiled against OpenAI.
Read the Complaint
You can read the full class action complaint in Lim v. OpenAI Global, LLC below.Sources
• Lim v. OpenAI Global, LLC, No. 3:26-cv-04063 (N.D. Cal.) — class action complaint, filed May 5, 2026 (embedded above)• Lim v. OpenAI Global, LLC, No. 3:26-cv-04063 (N.D. Cal.) — court docket, including the May 13, 2026 notice of voluntary dismissal and May 15, 2026 case termination
• Cal. Penal Code §§ 630–638 (California Invasion of Privacy Act)
• 18 U.S.C. § 2511 et seq. (Electronic Communications Privacy Act)
About This Page
This article summarizes a class action complaint filed in Lim v. OpenAI Global, LLC, No. 3:26-cv-04063, in the U.S. District Court for the Northern District of California. The allegations described here are the plaintiff's claims; they are unproven, OpenAI has not been found liable, and the case was voluntarily dismissed without prejudice on May 13, 2026. OpenClassActions.com is a consumer news and information site and is not a law firm or a party to this case. This page is general information, not legal advice.
For more class actions keep scrolling below.
Status
Voluntarily Dismissed Without Prejudice (Case Closed)
Case Title
Saje Lim v. OpenAI Global, LLC
Case Number
3:26-cv-04063
Court
U.S. District Court, Northern District of California
Date Filed
May 5, 2026
Voluntarily dismissed May 13, 2026; case terminated May 15, 2026
Plaintiff's Counsel
Bursor & Fisher, P.A.
Complaint
Read the Full Complaint (PDF)
