Google has quietly rewritten a key privacy promise. A policy update that took effect October 16 carves out an exception to Google's pledge that data from Gmail, Drive, and Google Photos would not be used for personalized ads — and the exception applies when users connect those services to Google's AI features.
The change was first spotted by Android Authority and reported in detail by The Verge on Thursday. It affects a growing set of users who have opted into Google's "Personal Intelligence" push, which lets AI Mode in Search and other Google tools read across a person's Gmail, Drive files, and photo library to deliver more tailored answers and recommendations. For more context on this story, see our ongoing AI news.
What Changed in the Policy
Under the previous version of the privacy policy, Google stated flatly that it would not use content from Drive, Gmail, or Google Photos to personalize ads. The revised policy preserves that promise — with a new escape hatch. As of October 16, the pledge carries the added line: "unless you choose to make data from these services available to other Google services that support personalized ads."
The updated policy text reads: "You may choose to allow certain Google services, including services like Search or Gemini Apps, to access data from other Google services you use, such as Google Photos and Workspace apps like Gmail and Drive, in order to provide you with a more personalized experience, subject to the policies and settings of the accessing service."
The critical sentence follows: "For example, Search may process some of the data that you allow Search to access to personalize content, including ads."
In plain terms: the moment a user wires Gmail, Drive, or Photos into AI-powered Search features, that same data flow can inform which ads they see elsewhere in Google's ecosystem.
How Your Gmail Data Can End Up Shaping Ads
The mechanism runs through Google's Personal Intelligence settings. Users who connect their apps give Search and Gemini-style tools permission to draw on personal content — an email confirmation, a saved document, a photo album — to answer questions like "when is my flight?" or "find the receipt for that laptop."
According to The Verge's report, anyone who has connected Drive, Gmail, or Google Photos to AI Mode and other Search tools "may now see ads targeted on the basis of that content." Google's own policy example illustrates the broader principle: search for "mountain bikes" in Search, and you may see ads for sports equipment on YouTube.
Google frames this as a choice, not a silent switch. The data bridge only exists for users who opted into the personalization features in the first place. But privacy watchers note the practical reality: the AI features are being promoted aggressively across Google's surfaces, and few users reading a consent prompt for "better, more personalized AI answers" would expect it to open the door to ad targeting based on their private emails and photos.
What Google Says Is Still Off-Limits
Google's guardrails, as described in the report, remain significant. Advertisers do not receive access to the underlying Gmail, Drive, or Photos content, nor to personally identifying details such as a user's name or email address. Google also says it will not serve ads based on sensitive categories — race, sexual orientation, religion, or health — regardless of what its systems can infer.
That last commitment matters given what those services contain. Medical documents in Drive, correspondence with religious organizations in Gmail, or health-related photos could, in principle, reveal sensitive attributes. Google's policy says ad targeting based on such categories stays off the table; the ads are shaped by interests and activity patterns rather than raw content exposure.
How to Opt Out
Users who want to sever the connection can do so in Search's personalization settings, where Google allows people to choose which apps are disconnected from AI features. Disconnecting Gmail, Drive, or Photos there removes the data bridge — and with it, per the policy language, the basis for ad personalization drawn from those services.
The Verge's guidance is straightforward: if you never connected these apps to AI Mode and Search's personal tools, the change does not affect you. If you did, and the new policy language troubles you, the settings page is the lever.
Why It Matters
The episode is a case study in how AI features are reshaping data flows that companies once pledged to firewall. Google spent years building the impression that Workspace content sat outside its ad machinery — a distinction that mattered to consumers, regulators, and enterprise customers choosing between Google Workspace and rivals.
Generative AI erodes that wall by design. Personalization is the product: an assistant that knows your calendar, your files, and your photos is more useful than one that does not. Google's answer to the resulting tension is consent-plus-exception — keep the old promise in the policy, then add a clause that lets the AI era route around it.
Regulators on both sides of the Atlantic have scrutinized exactly this pattern. Whether Google's framing — user-initiated, clearly documented, and walled off from advertiser access — satisfies them is an open question. For now, the practical takeaway for users is narrower: connecting your inbox to an AI assistant is no longer just a convenience decision. As of October 16, it is an advertising one too.
---
Stay Ahead of AIGet the latest AI news, analysis, and breakthroughs — all in one place.
Read more AI news →
