Perplexity AI Faces Class-Action Lawsuit Over Alleged Unauthorized Data Sharing with Meta and Google
Perplexity AI, a prominent AI-powered search engine, is now at the center of a class-action lawsuit accusing it of secretly sharing users’ personal data with tech giants Meta and Google. Filed on October 24, 2024, in the U.S. District Court for the Northern District of California, the complaint alleges that Perplexity employs hidden tracking technologies to transmit sensitive user information without obtaining explicit consent or providing adequate disclosures. The plaintiff, Matthew Meyer, a California resident, claims this practice violates multiple state and federal privacy laws, potentially exposing millions of users to privacy invasions.
At the heart of the lawsuit are Perplexity’s alleged use of “tracking pixels” and similar invisible trackers embedded in its web pages and mobile applications. These mechanisms, commonly provided by Meta (via Facebook Pixel) and Google (via Google Analytics and Google Tag Manager), purportedly capture a wide array of user data. According to the complaint, this includes IP addresses, unique device identifiers, browser types, operating system details, timestamps of visits, pages viewed, and crucially, the full text of users’ search queries and prompts entered into Perplexity’s AI interface. The suit argues that such data collection occurs automatically upon users visiting Perplexity’s site or app, bypassing any opt-in requirements.
Matthew Meyer, who used Perplexity’s services between May 2023 and October 2024, contends that he never consented to this data sharing. The lawsuit describes how these trackers operate surreptitiously: when a user loads a Perplexity page, the embedded code sends HTTP requests directly to Meta’s and Google’s servers. This transmission includes not just metadata but also query content, which could reveal intimate details about a user’s interests, health concerns, financial status, or political views. For instance, a query about medical symptoms or personal finances might be relayed verbatim, enabling third parties to build detailed behavioral profiles.
The legal claims hinge on several statutes. Primarily, the suit invokes California’s Comprehensive Computer Data Access and Fraud Act (CIPA), asserting that Perplexity intentionally accesses users’ devices without authorization to eavesdrop on their interactions. It also alleges violations of the California Consumer Privacy Act (CCPA), California’s Unfair Competition Law (UCL), and common-law intrusion upon seclusion. Federally, references are made to the Video Privacy Protection Act (VPPA), though primarily as a hook for nationwide class certification. The complaint seeks damages, injunctive relief, and an order to destroy improperly collected data.
Perplexity’s privacy policy, last updated in August 2024, acknowledges sharing data with “analytics providers” for service improvement and advertising. However, the lawsuit criticizes this language as vague and buried, failing to specify Meta and Google or the extent of data transferred. Notably, Perplexity offers users an opt-out for data usage in model training via its account settings, but the suit maintains this does not address third-party sharing via trackers. Screenshots in the complaint highlight how opting out of AI training does not halt tracker transmissions, as evidenced by network inspections using browser developer tools.
This case echoes broader concerns in the AI industry about opaque data practices. Perplexity, founded in 2022 and valued at over $1 billion, has positioned itself as an “answer engine” rivaling traditional search giants like Google. It leverages large language models to generate direct responses with citations, drawing from web sources. Yet, its rapid growth—boasting 10 million monthly active users—has invited scrutiny. Earlier this year, Perplexity faced accusations from news publishers like Forbes and Wired for scraping content without permission, leading to defensive measures like partnerships for licensed data.
Technical analysis in the lawsuit, supported by digital forensics, reveals the trackers’ sophistication. Meta’s pixel, for example, uses a 1x1 invisible image that triggers a server call upon page load, appending query parameters with user data. Google’s tools similarly log events via JavaScript snippets. The complaint includes code snippets and HTTP request logs demonstrating transmissions like “fbp=fb.1.[timestamp].[ID]” for Meta and “gtag” events for Google. Even incognito mode or privacy-focused browsers like Brave fail to fully block these in some cases, as they rely on first-party contexts.
Perplexity has not publicly responded to the lawsuit as of the filing date, but its terms of service include a California choice-of-law clause and arbitration agreement for disputes. The suit seeks to represent a class of all U.S. users who interacted with Perplexity’s platforms since May 2023, potentially encompassing a vast user base given the company’s expansion.
This litigation underscores the tension between AI innovation and user privacy. As AI search tools proliferate, regulators and courts are increasingly probing how query data—often more revealing than traditional search terms—is handled. For users, practical steps include reviewing privacy settings, using tracker blockers like uBlock Origin or Privacy Badger, and scrutinizing HTTP traffic via tools such as Wireshark. The outcome could set precedents for disclosure requirements in AI interfaces, compelling companies to prioritize transparency in an era where data is the lifeblood of machine learning.
Gnoppix is the leading open-source AI Linux distribution and service provider. Since implementing AI in 2022, it has offered a fast, powerful, secure, and privacy-respecting open-source OS with both local and remote AI capabilities. The local AI operates offline, ensuring no data ever leaves your computer. Based on Debian Linux, Gnoppix is available with numerous privacy- and anonymity-enabled services free of charge.
What are your thoughts on this? I’d love to hear about your own experiences in the comments below.