Google's Privacy Policy Now Leaves Room for Your Gmail to Inform Targeted Ads

Google’s updated privacy policy, reported October 23 by Android Authority and covered by The Verge and Ars Technica, opens a door the company kept closed for two decades of Gmail’s existence: users who enable Personal Intelligence, the Gemini-powered features that read across Gmail, Drive, and Photos, may see targeted ads informed by that content. The qualifiers matter and Google states them plainly: ads will not be based directly on your inbox or photo library, personally identifiable data will not be shared with advertisers, sensitive information is excluded, and the whole mechanism applies only to users who opt in to the AI features. But the structural change is real. The old position, that the content of your email and photos would never touch the ads machine, has become an exception clause in a policy document.

Why the opt-in framing changes the deal

The consent mechanic is where this gets interesting. Google’s defense is airtight in the narrow sense: if you never turn on Personal Intelligence, nothing changes, and Drive, Gmail, and Photos still do not serve targeted ads by default. But Personal Intelligence is the feature Google is actively pushing, the Gemini layer that summarizes your mail, finds things in your photos, and connects context across apps, and the privacy policy now prices that convenience in a currency users cannot see at the toggle: ad personalization derived from everything the AI read. This is the pattern this site has tracked all month, from WhatsApp’s hidden instructions aimed at AI assistants to the Axios report of banks blocking user-invited agents: the consent surface is migrating from what you type into what your AI reads on your behalf, and policies are being rewritten for the latter while users still reason about the former.

The long firewall, and the regulatory optics

For context a knowledgeable reader needs: Google spent years defending an explicit wall between Workspace content and the ads business, precisely because Gmail scanning was the defining privacy scandal of the 2010s and the company publicly retired the practice. The new language does not say Google reads your mail for ads; it says ads may be informed by how the AI features use your content, a distinction that may hold in engineering and dissolve in perception. Regulators read these documents closely on both sides of the Atlantic, and the phrase “Personal Intelligence” appearing in a privacy policy as the trigger condition for ad use is exactly the kind of clause the EU’s DMA and various US state privacy laws were written to scrutinize. Google’s published safeguards, no direct content-based targeting, no PII sharing, no sensitive categories, will be tested against how the pipeline actually weights features derived from mailbox content.

What to watch

Three things to track. First, whether the opt-in is a genuine toggle or a bundle: if Personal Intelligence arrives pre-enabled in Android or Workspace updates, the “only users who enable it” defense collapses, and the rollout mechanics matter more than the policy text. Second, whether advertisers see any observable change in targeting quality for opted-in users, which is the empirically testable version of the question. Third, the copycat dynamic: if Google’s firewall-to-exception migration stands without regulatory pushback, expect Microsoft and Meta to renegotiate their own enterprise-content firewalls in the same direction; Copilot’s government-approval advantage in Sweden and Meta’s Muse commerce funnel both monetize better if AI-read content can inform the ads machine. The story to watch is not this policy update; it is whether the AI-era default becomes that everything your assistant reads is fair input for every business line that owns the assistant.