Here is a fun experiment: go to Perplexity AI, ask it something deeply personal about your finances, your health, or your marriage. Now imagine Meta and Google got a copy of that conversation. According to a class-action lawsuit filed in San Francisco federal court, that is exactly what has been happening.

Bloomberg reported on Tuesday that a Utah man, identified as John Doe, filed a proposed class-action lawsuit alleging that Perplexity embeds hidden tracking software into its search engine code. The trackers allegedly activate the moment a user logs in, giving Meta and Google full access to every conversation between the user and the AI.

The worst part? It happens even when you use Incognito mode. The feature Perplexity offers that is supposed to mean your queries stay private. According to the complaint, that toggle does absolutely nothing to stop the data from flowing to two of the largest advertising companies on the planet.

The Privacy Paradox That Should Terrify You

Think about what Perplexity is. It is an AI search engine. People use it for the kinds of questions they would never type into Google because Google already knows too much. Medical symptoms. Legal questions. Financial planning. Relationship advice. The entire value proposition is that it is supposed to be smarter and more private than Google Search.

If this lawsuit is accurate, Perplexity was taking those private conversations and routing them straight back to Google anyway. That is not just a privacy violation. That is a betrayal of the core product promise.

The plaintiff claims he shared details about his family finances, tax obligations, and personal investment strategies with Perplexity. The complaint alleges Meta and Google can exploit that data for targeted advertising and even resell it to third parties.

Perplexity Cannot Catch a Break

This is not Perplexity's first legal headache. Amazon is currently suing the company over its agentic shopping feature, alleging that Perplexity covertly accessed Amazon customer accounts while disguising automated activity as human browsing. The company is building a reputation for moving fast and asking forgiveness later. Except the lawsuits keep suggesting the forgiveness is not coming.

Perplexity spokesperson Jesse Dwyer told Bloomberg they have not been served any lawsuit matching this description. Meta pointed to a help page saying advertisers are not supposed to send them sensitive information. Neither response inspires confidence.

What Happens Next

If this class-action gains traction, every AI search engine will face the same question: what are you doing with the conversations people trust you with? Perplexity raised $500 million at a $9 billion valuation. That valuation assumed users would trust the product. Trust, once broken, does not come back with a press statement.

The bigger story is not Perplexity. It is the entire AI industry. If the companies building AI search tools are quietly feeding your data to the same advertising giants you were trying to escape, the privacy promise of AI is a lie. And consumers are starting to lawyer up.