A federal class action filed March 31 accuses Perplexity AI, Google, and Meta of secretly collecting AI conversation data, including health and finance queries, via embedded trackers.
A class action complaint filed on March 31, 2026, in the United States District Court for the Northern District of California accuses Perplexity AI, Inc. of secretly embedding advertising trackers inside its AI search engine - then using those trackers to forward users' private conversations to Google and Meta without their knowledge or consent. The case, Doe v. Perplexity AI, Inc., Meta Platforms, Inc., and Google, LLC, Case No. 3:26-cv-02803, was assigned to the San Francisco Division. Both Google and Meta are named as co-defendants.
The complaint is 140 pages long and was filed on behalf of a plaintiff identified as John Doe, a citizen of Utah, who used Perplexity's AI platform both as a guest and as a registered account holder. The law firm Ahmad, Zavitsanos & Mensing, PLLC, based in Houston, Texas, represents the plaintiff and the proposed class.
According to the complaint, Perplexity embedded the Facebook Meta Pixel, Google Ads, Google DoubleClick, Google Firebase Analytics, and Google Firebase code directly inside its AI platform and its mobile application. These tools are standard components of digital advertising infrastructure, but the complaint argues their deployment inside an AI conversation engine crossed a legal line.
The moment a user visited www.perplexity.ai for the first time, according to the complaint, the website's server returned source code that "automatically and instantaneously downloads tracking pixels on its users' browsers, which then hijack those browsers, and instruct them to begin sharing all future communications between users and Perplexity with Meta and Google." This happened before the user clicked a single button or entered a single prompt.
The technologies involved work through several mechanisms. The Meta Pixel is a script embedded in a website's source code that redirects information about user actions to Meta's servers. A cookie named "c_user" directly links activity to a user's Facebook profile if they are logged in on the same browser. A second cookie, called the "fr" cookie, goes further - it tracks both Facebook users and non-Facebook users, allowing Meta to build behavioral profiles regardless of whether the person holds an account on Meta's platforms. The complaint explains that via the "fr" cookie, "Meta can collect information such as visited websites, button clicks, interactions with content and more, which contributes to building detailed user profiles."
A separate tool - Meta's Conversions API - transmitted web events and user interactions directly from Perplexity's servers to Meta's servers, bypassing the browser entirely. This server-to-server pathway does not depend on browser cookies and is therefore harder for users to detect or block.
On the Google side, the complaint describes Google DoubleClick as "an advertising tracking tool designed to collect information about user behavior on websites for the purpose of improving the targeting and effectiveness of advertising campaigns." DoubleClick was originally an independent company acquired by Google and has since been integrated into the Google Marketing Platform. Google Analytics and Google Firebase code ran separately inside Perplexity's mobile application, the complaint states, together creating "a vast and comprehensive amount of its users' data being shared with Google."
Collectively, the complaint refers to these tools as the "Tracking Technologies." According to the filing, they captured at minimum: the user's opening prompt in each conversation, even when subscribers were in "Incognito Mode"; follow-up questions that users clicked on from options presented within the dialogue; sign-up email addresses; and every conversation URL - which for non-subscribed users contained both the initial prompt and a link through which third parties could access the entire conversation.
One of the complaint's more specific allegations concerns Perplexity's own "Incognito" mode. Perplexity offers this feature to account holders as a way of conversing without the platform retaining history. According to the complaint, the feature did nothing to prevent the underlying tracking technologies from operating. "Even paid users who turned on the 'Incognito' feature still had their conversations shared with Meta and Google, along with their email addresses and other identifiers that allowed Meta and Google to personally identify them," the filing states.
Nothing on Perplexity's website, according to the complaint, warned users that their conversations would be shared with third parties. No specific disclosure was made to subscribed users that Incognito mode did not function as protection against that sharing.
The named plaintiff, John Doe, described a range of personal financial topics he researched using Perplexity's AI platform - including when he and his spouse could begin withdrawing social security, how to convert savings from a taxable brokerage account into a Roth IRA, potential stock trades and price analysis, including for publicly traded cannabis companies, and the rules governing Canadian companies' share buybacks in a single year. He also sought tax advice and legal information.
According to the complaint, Doe believed "that these communications with Perplexity's AI Machine were private and that Perplexity would not share his communications with third parties like Meta and Google." He was, the filing states, "dismayed to discover that complete and partial transcripts of his communications with Perplexity were shared with Meta and Google every time that he interacted with Perplexity's AI Machine."
The complaint does not treat the tracking as an accident or a configuration error. It frames it as a deliberate business arrangement. According to the filing, "In return for sharing all this sensitive data with Meta and Google, Perplexity received advertising and analytics benefits that it used for its own benefit."
Perplexity was founded in San Francisco in 2022. According to the complaint, it has raised more than $1.5 billion in funding, including from Nvidia and Jeff Bezos. In September 2025, the company secured $200 million in a funding round that raised its valuation to $20 billion.
The complaint argues that Meta and Google were fully aware of what they were receiving. According to the filing, "Meta and Google knew that the data collected and received from Perplexity's AI Machine included intimate personal health and financial data - but they did nothing to stop Perplexity from sharing this data because it is vital to their business models."
The case asserts violations of several federal and California statutes. These include the California Invasion of Privacy Act (CIPA), CAL. PENAL CODE sections 630 et seq.; the Comprehensive Computer Data Access and Fraud Act(CDAFA), CAL. PENAL CODE section 502; the California Constitution, Article 1, section 1; and the Electronic Communications Privacy Act (ECPA), 18 U.S.C. section 2511 et seq.
Disclosing the contents of private communications without consent is characterized in the complaint as "a crime in California" under CAL. PENAL CODE section 631. The filing also cites a January 2026 ruling from the Northern District of California - Krzyzek v. OpenX Techs., Inc. - in which that court found that "engaging in unauthorized tracking and data collection, allowing it to compile detailed profile of each plaintiff's online web browsing activity - including highly sensitive browsing activity - tied to their email addresses and other personal identifiers is actionable."
According to the complaint, the aggregate amount in controversy exceeds $5 million, and the class is estimated to include more than 100 putative members - a threshold that triggers jurisdiction under the Class Action Fairness Act of 2005. The plaintiff seeks statutory damages of at least $5,000 per violation, an injunction stopping Perplexity from further unauthorized disclosures, attorney's fees and costs, and any other relief the court finds appropriate.
The case encompasses all individuals in the United States whose conversational dialogues with Perplexity's AI platform were shared with Meta and Google without their consent. The complaint explicitly includes guest users - people who never registered for an account.
As of the date the complaint was filed, Perplexity and Meta had not issued official statements in response. Google told Ars Technica - which first identified the court document - that "Businesses manage the data they collect and are responsible for informing users about it," according to the PCMag report covering the complaint. Google added: "By default, data sent to Google Analytics for measurement does not identify individuals, we have strict policies against advertising based on sensitive information, and we don't sell personal information."
The filing arrives against a backdrop of growing legal pressure on Perplexity. In November 2025, Amazon sued the company over its Comet browser's agentic shopping behavior, alleging the browser disguised itself as Google Chrome while enabling AI agents to access private customer accounts and make purchases. A court subsequently ordered Perplexity to block its AI agents from placing orders on Amazon without permission, though the company indicated it would contest that ruling.
Earlier, Reddit filed a federal lawsuit in October 2025, naming Perplexity AI among several defendants for allegedly circumventing technological controls to access platform content. Perplexity denied training models on Reddit's content, while Cloudflare documented the company using stealth crawlers to evade protections.
Questions about Perplexity's approach to user data predate this lawsuit. In May 2025, CEO Aravind Srinivas disclosed during a podcast interview that Perplexity intended to track user activities across the internet through its Comet browser to build comprehensive profiles for advertising purposes. PPC Land reported on that strategy, noting that the approach mirrored the data collection infrastructure pioneered by the very platforms Perplexity was positioned to challenge.
Perplexity has been building out its advertising business since late 2024. The company launched its first ad formats in November 2024 - sponsored follow-up questions and paid media adjacent to answers - with partners including Whole Foods Market, Universal McCann, and PMG. An earlier pitch deck from August 2024 described the platform's anticipated CPM rates at over $50, with target categories including health, technology, and finance.
The lawsuit raises a question that sits at the heart of how AI platforms can be monetized. Perplexity's public statements to advertisers emphasized that personal information would not be shared with advertisers and that answer content would not be influenced by advertising. The complaint, if its allegations are proven, would sit in direct tension with that framing - the data appears to have flowed not to the advertisers themselves but upstream to the infrastructure providers who power targeting across the wider web.
The type of data at issue matters enormously for ad tech. Health and financial queries made to an AI conversation engine are not generic browsing signals. They are intent-rich, highly sensitive disclosures made by users who believed they were speaking to a private assistant. The complaint's liver cancer example is illustrative: a user asking about treatment options could subsequently receive advertising for oncology clinics, alternative therapies, or pain medications - targeted on the basis of what they confided in a tool marketed as a research aid.
This dynamic is not unfamiliar to regulators. In Europe, Meta's Business Tools tracking infrastructure has faced repeated adverse rulings. German courts have consistently found the Meta Pixel and Conversions API violate GDPR when used to collect data without valid consent. A German court awarded one user €5,000 in compensation in July 2025, specifically for violations involving Meta's Business Tools tracking infrastructure. The Doe v. Perplexity complaint cites CIPA's $5,000-per-violation damages figure, mirroring that scale.
In the United States, the trajectory of privacy enforcement has been toward higher penalties and broader class definitions. A federal jury found Meta violated CIPA in August 2025 by collecting sensitive health data from users of the Flo period-tracking app without their consent - a case involving Meta's SDK embedded inside a third-party application, structurally similar to the role the Meta Pixel allegedly played inside Perplexity's platform. California entered a final $50 million judgment against Meta in March 2026 over Facebook user data shared with third-party developers.
The Doe complaint names all three parties jointly: Perplexity as the deployer of the tracking technologies, and Google and Meta as the recipients who, the filing contends, actively chose to receive and exploit the data. That tripartite liability theory - if accepted by the court - would be significant for any publisher, app developer, or AI platform currently embedding Google or Meta measurement tools while handling sensitive user queries.
Who: The plaintiff is John Doe, a Utah resident and Perplexity user, filing on behalf of a proposed class of all United States users whose conversations with Perplexity's AI platform were shared with Meta and Google without consent. The defendants are Perplexity AI, Inc., Meta Platforms, Inc., and Google, LLC.
What: A 140-page federal class action complaint alleges that Perplexity secretly embedded the Meta Pixel, Conversions API, Google Ads, Google DoubleClick, Google Firebase, and Google Analytics inside its AI search platform and mobile application, forwarding users' private conversations - including sensitive health, financial, and legal queries - to Meta and Google. The complaint seeks at least $5,000 in statutory damages per violation, injunctive relief, and attorney's fees.
When: The complaint was filed on March 31, 2026. The alleged tracking is claimed to have occurred throughout the period during which users interacted with Perplexity's platform, including while using the company's "Incognito" mode.
Where: The case was filed in the United States District Court for the Northern District of California, San Francisco Division, Case No. 3:26-cv-02803. Perplexity's principal place of business is listed as 115 Sansome Street, Suite 900, San Francisco, California.
Why: The complaint argues that Perplexity received advertising and analytics benefits in exchange for sharing user data, while Meta and Google chose to receive and exploit that data as a core part of their advertising business models. The plaintiff contends that no reasonable person would have anticipated that private AI conversations - about cancer treatments, retirement planning, or stock analysis - would be forwarded to the world's largest advertising platforms without disclosure or consent.
