A federal class-action lawsuit filed on March 31, 2026 in San Francisco alleges that Perplexity AI embedded tracking software in its platform that transmitted user conversations to Meta and Google, without user knowledge or consent.
The 135-page complaint was filed by an anonymous plaintiff identified as John Doe. It claims that trackers are downloaded onto users' devices as soon as they log into Perplexity's homepage, giving Meta and Google access to the full content of user conversations: both the questions users ask and the AI's responses.
The complaint names Perplexity, Meta, and Google as defendants, alleging violations of California privacy law, federal computer privacy statutes, and fraud laws.
Perplexity said it had not been served the lawsuit. "We have not been served any lawsuit that matches this description so we are unable to verify its existence or claims," said Jesse Dwyer, the company's chief communications officer. Meta pointed to its help page stating that advertisers are prohibited from sending it sensitive information. Google did not respond to requests for comment.
The allegations are unproven and the case is in its earliest stages, but the questions it raises about how AI assistants handle private conversations are significant regardless of how the litigation resolves.
What the Lawsuit Actually Alleges
The complaint's central claim is that Perplexity embedded "undetectable" tracking software that automatically transmits user conversations to Meta, Google, and potentially other third parties, before those conversations are even seen by Perplexity itself.
According to the filing, the trackers harvest users' email addresses, Facebook IDs, IP addresses, and device and browser information. Once Meta and Google have that data, the complaint argues, they can pair it with a user's name and address to specifically identify them and target them with advertising.
The suit alleges that for free users, initial prompts are shared with a URL structure that gives third parties access to the entire conversation thread. Follow-up questions that users click on are also transmitted.
What the Complaint Says Was Shared
| Data Category | Alleged Transmission |
|---|---|
| Full conversation content | Both user prompts and AI responses |
| Email address | Linked to conversation data |
| Facebook / Google ID | Used to identify users across platforms |
| IP address | Enables location and identity matching |
| Device and browser info | Enables cross-device tracking |
| Conversation URL (free users) | Allows third-party access to full thread |
The plaintiff also alleges that the trackers remained active in Perplexity's Incognito mode, which the company describes as creating "anonymous threads" that "won't save to your history and expire after 24 hours."
If the allegation is accurate, that privacy promise is functionally empty for free users.
The complaint includes a detail that helps explain why this kind of tracking is particularly sensitive for an AI platform. It cites studies finding that "many users specifically and intentionally turn to AI Machines for issues they are reluctant to discuss with other humans, including such things as relationship advice, companionship, and sexual or identity exploration."
The named plaintiff is a Utah man who used Perplexity to research Social Security timing, Roth IRA conversion strategies, and potential cannabis company investments. He believed those conversations were private.
Who Is Included in the Lawsuit
The complaint seeks to certify two classes.
The first is a nationwide class of all users who chatted with Perplexity between December 7, 2022 and February 4, 2026, excluding subscribers with paid Pro or Max subscriptions.
The second is a California-only subclass covering the same period and the same exclusion.
The reason paid subscribers are excluded is not explained in detail, but a footnote in the complaint notes that paid subscribers operate under different terms and conditions. The plaintiff holds a free account.
The Context: This Is Not Perplexity's First Legal Problem
The lawsuit lands at a moment when Perplexity is already facing scrutiny on multiple fronts.
The company was valued at $20 billion in September 2025 when it raised $200 million from Institutional Venture Partners and others. That valuation, combined with the volume of sensitive queries it processes, makes it an increasingly significant target for regulators and plaintiffs.
Perplexity has faced prior legal challenges in 2026 related to its AI agent and e-commerce features. A search of recent litigation, according to Bay City News Service reporting, did not identify other suits against major AI developers based on the specific tracking allegations in this complaint. The wiretap-style framing, describing the trackers as "browser-based wiretap technology," may prove significant as courts decide which legal framework applies to AI conversation data.
What Perplexity's Privacy Policy Actually Says
Perplexity's official help center states that it "does not sell, trade, or share your personal information with third parties, except as outlined in our policy."
The policy's listed exceptions include sharing with service providers who help operate the platform, such as payment processors and customer support tools, and with "advertising and analytics partners who deliver personalized ads across the platform."
That last category is where the gap between the stated policy and the lawsuit's allegations becomes legally significant. Sharing data with advertising and analytics partners for purposes of targeted advertising is not the same as selling data, but it is precisely the kind of sharing that the plaintiffs allege was undisclosed and non-consensual.
The privacy policy also acknowledges that "most of the above information" may be used to "monitor and analyze trends" and to "enhance the safety and security of our services."
For enterprise users, Perplexity offers substantially stronger protections. Enterprise Pro and Max accounts operate under zero data retention and zero data training agreements, meaning query content is not used to train models and is not retained beyond the session.
The complaint specifically excludes paid subscribers from the class, which suggests the tracking mechanism the lawsuit targets is either absent or materially different for paying accounts.
Free vs. Paid Account Privacy Comparison
| Feature | Free Account | Pro / Max Account | Enterprise |
|---|---|---|---|
| AI training on your queries | Enabled by default (opt-out available) | User-controlled | Never |
| Data retained after session | Yes, by default | User-controlled | Strict ZDR policy |
| Third-party ad tracker exposure | Alleged in lawsuit | Excluded from complaint | Different terms |
| Incognito mode privacy | Alleged to be ineffective | N/A | Incognito can be enforced by admin |
| History saved | Yes, unless Incognito | User-controlled | Admin-controlled |
How to Protect Yourself Right Now
The lawsuit is unproven and the court has not yet ruled on whether the tracking occurred as alleged. The practical protective steps below are worth taking regardless of how the litigation resolves, because the data practices described in the complaint, including advertising partner data sharing, are not disputed in Perplexity's own policy language.
Step 1: Opt Out of Targeted Advertising in Perplexity Settings
Perplexity's privacy policy includes an opt-out for online targeted advertising. You can access it through the cookie settings link on the homepage, or through the "Do Not Share" link. The company also states it recognizes Global Privacy Control (GPC) signals from browsers as a valid opt-out where required by law. Note that GPC applies only to the browser you set it on, not across devices.
Step 2: Turn Off AI Data Retention
For free, Pro, and Max users, AI data retention is enabled by default. Perplexity's help center confirms you can turn this off:
- Log into your Perplexity account.
- Go to Account Settings.
- Find the AI Data Retention toggle and disable it.
This stops your query data from being used to train Perplexity's models. It does not necessarily stop data from being processed by third-party trackers, which is the core of the lawsuit's allegation.
Step 3: Stop Sharing Sensitive Information in Free Tier Queries
The plaintiff in this lawsuit shared financial planning details, tax strategies, and investment research in conversations he believed were private. Regardless of whether the lawsuit succeeds, the safer practice is to treat any free-tier AI assistant query as potentially visible to third parties.
Queries that should not go into a free AI platform include:
- Tax, financial, or investment details tied to your identity
- Health or medical information
- Legal situations or ongoing disputes
- Login credentials or account information
- Personal relationships or identity-related questions
If you need to research sensitive topics, use a paid tier with explicit privacy terms, or rephrase queries to remove identifying details.
Step 4: Block Third-Party Trackers at the Browser Level
The complaint describes the tracking mechanism as browser-based. Several tools can limit this exposure regardless of the platform you are using:
- uBlock Origin blocks a wide range of third-party advertising and analytics trackers, including Meta Pixel and Google Analytics.
- Privacy Badger (Electronic Frontier Foundation) automatically blocks trackers based on behavior.
- Firefox with Enhanced Tracking Protection set to "Strict" blocks most known third-party trackers by default.
- Brave browser blocks trackers and third-party cookies by default with no configuration required.
Using a browser-level tracker blocker is one of the most effective things you can do, because it operates before the website's JavaScript has a chance to transmit your data.
Step 5: Upgrade to a Paid Plan or Use a More Privacy-Focused Alternative
The lawsuit's class excludes paid Pro and Max subscribers, which suggests the tracking mechanism operates differently for accounts with separate terms. If you use Perplexity regularly for sensitive research, upgrading reduces but may not eliminate your data exposure, and you should review the specific terms attached to paid accounts before assuming your data is fully protected.
Alternatives with more explicit privacy positioning include Kagi (a paid search engine with no ad tracking and no data resale) and Claude.ai (which has clearer data handling policies and does not currently operate an advertising business).
Why This Matters Beyond Perplexity
The lawsuit's deeper implication is not specific to one company.
AI assistants are increasingly being used as confessional and advisory tools. People tell them things they would not tell a doctor, a lawyer, or a family member. The social architecture of conversational AI encourages this: it is designed to feel like a private dialogue.
If third-party advertising trackers are running inside that dialogue, the privacy model that makes conversational AI feel trustworthy is built on a false premise. Users making financial, health, or personal decisions based on AI advice are not just sharing query data with the AI company.
They may be sharing it with the advertising infrastructure of the two largest platforms on the internet, without knowing it.
The outcome of this lawsuit will not resolve that question for the industry as a whole, but it will establish whether the specific tracking practice alleged here violated the law. That precedent will matter for every AI company currently deciding how to monetize its free user base.
Wrap up
A federal class-action lawsuit alleges Perplexity transmitted free users' conversations to Meta and Google through hidden trackers that operated even in Incognito mode. Perplexity says it has not been served the suit, and Meta and Google deny wrongdoing. The allegations are unproven.
What is not in dispute is that Perplexity's privacy policy permits data sharing with advertising and analytics partners and that AI data retention is enabled by default for free users. The company's Incognito mode does not carry the same privacy protections as end-to-end encrypted communication.
The practical advice is the same whether or not the lawsuit succeeds: do not share sensitive personal information in free AI assistant queries, turn off data retention where the platform allows it, use a browser-level tracker blocker, and understand that free AI products are not the same as private ones.
Frequently Asked Questions
What is the Perplexity lawsuit about?
A federal class-action complaint filed March 31, 2026 in San Francisco alleges that Perplexity AI embedded hidden tracking software that transmitted user conversations to Meta and Google without user consent. The suit claims trackers download at login and give third parties access to the full content of conversations, including both user queries and AI responses.
Does Perplexity's Incognito mode protect you?
The lawsuit alleges it does not. The complaint claims that trackers remained active even when users enabled Perplexity's Incognito mode, a feature described by the company as creating anonymous threads that do not save to history. This allegation is unproven, but users should not assume Incognito mode provides meaningful privacy protection from third-party trackers.
Does Perplexity sell user data?
Perplexity's official policy states it does not sell user data. However, the policy permits data sharing with "advertising and analytics partners who deliver personalized ads." The lawsuit alleges this data sharing occurred without disclosure or consent, which is the basis for the legal claim.
Who is included in the Perplexity class-action?
The lawsuit seeks to certify a nationwide class of users who chatted with Perplexity between December 7, 2022 and February 4, 2026, excluding paid Pro and Max subscribers. A California-only subclass is also proposed under state privacy law.
How can I opt out of tracking in Perplexity?
Perplexity offers an opt-out for targeted advertising through its cookie settings or the "Do Not Share" link on the homepage. It also recognizes Global Privacy Control browser signals. Separately, you can disable AI data retention in your account settings, though this addresses model training rather than third-party tracker transmission. Using a browser-level tracker blocker such as uBlock Origin provides an additional layer of protection.
Are paid Perplexity users affected?
The lawsuit specifically excludes paid Pro and Max subscribers, noting they operate under different terms. Perplexity's Enterprise tier also includes explicit zero data retention protections. However, no paid plan should be assumed to eliminate all data exposure without reading the specific terms that apply.
Related Articles
Dani
Mark
Dani
