Perplexity AI Was Secretly Sharing Your Searches With Meta and Google

What the lawsuit claims

On April 1, 2026, a Utah man filed a class-action lawsuit against Perplexity AI in the US District Court for the Northern District of California. The core allegation is simple: when you log into Perplexity's search engine, tracking software is automatically downloaded onto your device that transmits your conversations to Meta and Google.

According to the complaint, these trackers activate at login and operate continuously, forwarding sensitive chat data for targeted advertising and resale. The tracking persists even in Incognito mode, which most people assume prevents this kind of data collection.

Perplexity's spokesperson responded by saying they 'have not been served any lawsuit that matches this description.' Notably, they didn't deny the tracking practices described in the complaint.

Why AI searches are more revealing than web searches

There's a reason this matters more than regular search tracking. When you use Google, you type keywords. When you use an AI search engine like Perplexity, you have conversations. You provide context, ask follow-up questions, share specifics about your situation.

Think about what people ask AI assistants: 'How do I handle these photos from my ex?''What's the best way to share baby photos with family without social media?' 'Can someone find my home address from the GPS data in my photos?' These aren't keyword searches - they're intimate disclosures that paint a detailed picture of your life, concerns, and vulnerabilities.

If these conversations are being forwarded to advertising networks, the ad targeting implications are significant. Meta and Google could build profiles based on what you're worried about, interested in, or struggling with - information you shared thinking it was private.

Incognito mode doesn't protect you here

One of the most damaging claims in the lawsuit is that the tracking works in Incognito mode. Most people use Incognito with the assumption that their browsing activity isn't being recorded or shared. But Incognito mode only prevents your browser from saving local history - it doesn't stop websites from downloading trackers onto your device.

If Perplexity's tracker downloads at the point of login, then Incognito mode offers zero protection. You're logged into the service, which identifies you. The tracker knows who you are and what you're asking about. The only thing Incognito prevents is your browser from remembering the session locally.

This is a common misunderstanding that platforms exploit. 'Private browsing' sounds like privacy, but it's really just local cleanup. The server side sees everything.

What Meta and Google do with this data

Meta and Google run the world's two largest digital advertising networks. Both companies build detailed profiles of users based on every data point they can collect - browsing history, app usage, location, purchases, and any third-party data they receive from partners.

Meta recently updated its privacy policy to explicitly state that data from user interactions with its AI services will be used for targeted ads. This includes 'prompts that can include questions, messages, media and other information.' If Perplexity's trackers are sending conversation data to Meta, it slots directly into this advertising pipeline.

Google's advertising model works similarly. Any data Google receives about your interests, concerns, and behavior gets fed into its ad targeting system. When you search for information about photo privacy on Perplexity, you might see ads for cloud storage, VPNs, or camera products across Google's network - because they know what you were asking about.

This is part of a bigger pattern

The Perplexity case isn't happening in isolation. It's part of a growing trend where AI companies monetize user interactions in ways their users don't expect.

• Meta updated its privacy policy to use AI chat data for ad targeting, prompting EPIC to call for FTC intervention
• OkCupid shared 3 million user photos with an AI company without consent, leading to an FTC settlement
• Tinder is testing features that scan your device's camera roll using AI
• Multiple AI photo editors were caught leaking user photos earlier this year

The pattern is consistent: companies promise private, helpful AI interactions, then monetize the data those interactions generate. The value proposition is 'use our smart tool,' but the business model is 'sell what you tell us.'

What you can actually do

Be careful what you share with AI tools. Treat every AI interaction as potentially public. Don't share photos directly with AI assistants, don't ask questions that reveal sensitive personal details, and don't assume that because a service feels like a private conversation, it is one.

Use browser extensions that block trackers. Tools like uBlock Origin, Privacy Badger, and Firefox's built-in Enhanced Tracking Protection can block many third-party trackers. They're not perfect, but they're better than nothing.

Separate your tools by trust level. Use privacy-focused tools for sensitive queries. For personal photo sharing, use a platform that doesn't feed your data into advertising networks. Keep your photo library, your personal conversations, and your AI tools in separate ecosystems.

What this means for your photos

You might think a lawsuit about an AI search engine has nothing to do with your photos. But it does. AI services are increasingly interconnected. If you use Perplexity to search for 'best way to organize family photos' or 'how to share travel photos privately,' and that data gets sent to Meta and Google, it informs how those companies target you with ads - and what they know about your digital life.

More importantly, it shows that the companies building AI tools don't always treat your interactions as confidential. If an AI search engine might share your queries with ad networks, an AI photo editor might share your images with training pipelines. The assumption of privacy is the thing that keeps breaking.

The safest approach is compartmentalization. Keep your photos in a service that exists to protect them, not to monetize them. Viallo stores your photos on EU servers, doesn't process them with AI, and never shares data with third parties or advertising networks.

Frequently asked questions