OpenAI Is Being Sued For Sending Your ChatGPT Questions To Facebook And Google

The complaint names a single plaintiff, California resident Amargo Couture, who says she used ChatGPT throughout 2025 and 2026 to research health questions, financial decisions and other personal matters. She is suing on behalf of every United States resident who typed a query into ChatGPT.com while the tracking code was active. The claim at the center of the case: OpenAI embedded Meta's Facebook Pixel and Google Analytics into the ChatGPT web interface, and those tools quietly copied conversation data out to two advertising companies OpenAI never asked users about.

The complaint, filed May 13 in the U.S. District Court for the Southern District of California, alleges the two tracking systems captured different slices of the same conversations. For Meta, the filing says the data included query text alongside a Facebook identifier cookie carrying the user's unencrypted Facebook ID, a pairing that would let Meta connect a person's private ChatGPT questions to their real Facebook profile. For Google, the complaint alleges Google Analytics intercepted IP addresses, device identifiers, the standard _ga cookie, the Secure-3PSID cookie and hashed login email addresses.

The Legal Theory

Couture's suit leans on three separate bodies of law: the federal Electronic Communications Privacy Act, California's Invasion of Privacy Act, and California's state constitutional privacy protections. All three exist because federal wiretap law was written for phone calls and has spent decades being stretched to cover new ways of intercepting communications, and CIPA in particular has become the go-to claim in a wave of California suits against companies accused of embedding third-party tracking scripts inside anything a user might reasonably expect to be a private exchange.

This was not an isolated filing. At least two related complaints landed in California federal courts during the same month, one in the Northern District of California and this one in the Southern District, both aimed at the same underlying allegation: that OpenAI wired third-party ad and analytics code into a product built around people typing sensitive, personal, often deeply private questions into a text box.

The Product ChatGPT Is Sold As

The irony the complaint leans on is not subtle. OpenAI and its retail partners have spent two years marketing ChatGPT as a place people can bring exactly the kind of questions Couture says she asked it: symptoms she was worried about, financial decisions she was weighing, personal matters she was not ready to discuss with another person. That pitch only works if the box is actually private. A user asking ChatGPT about a medical symptom is not consenting to Meta learning about it through a Facebook Pixel, and the complaint argues that no reasonable user reading OpenAI's privacy materials would expect their raw conversation topics to be routed to two of the largest ad-targeting companies on the internet.

Where The Case Stands

OpenAI has not yet filed a formal response to the complaint. No class has been certified, no discovery has produced internal documents confirming why the tracking code was placed where it was, and no court has ruled on whether the alleged conduct meets the legal bar for a wiretap or invasion-of-privacy claim. This is a filed accusation, not an adjudicated fact, and it will likely take months before a judge rules on OpenAI's expected motion to dismiss. What is already true, and does not depend on the outcome, is that ChatGPT.com carried Meta and Google tracking infrastructure while carrying the label of a place to ask your most private questions. Whether a court eventually calls that a wiretap violation or a disclosed and permitted analytics practice, the company built its trust on the premise that the two would never look the same.