Google has quietly revised its privacy policy to let content from Gmail, Drive and Google Photos shape personalized ads — but only for users who opt into its Personal Intelligence feature. The carve-out took effect October 16, 2026, and rewrites a promise Google had kept word-for-word since July 1, 2021.
The original pledge was categorical: “We don’t show you personalized ads based on your content from Drive, Gmail, or Photos.” The new version keeps that sentence but adds a condition: “unless you choose to make data from these services available to other Google services that support personalized ads.”
What actually changed
The editing is precise and worth understanding. Google’s policy archive shows three key rewrites in the advertising section:
- The personalized ads sentence now covers “interests, activity, or personalized content from Google services.” The word “or” matters — it makes AI-generated personal content a third, separate basis for ad personalization.
- A new “personalize content” note names Personal Intelligence, Google’s opt-in feature that lets Search and Gemini read your emails, photos and files to give more personal answers.
- The note’s hotel booking example links search results to the same inference that can now shape ads.
Advertisers still don’t get inbox access. Google’s help page says ads “will not be based directly” on Gmail inboxes or Photos — they’ll be shaped by “summaries, excerpts, generated media, and inferences” from that content.
Why it matters for performance marketers
For five years, this data was formally excluded from Google’s ad systems. Now it’s a potential signal inside the auction.
Think about the use case Google itself describes: a user asks Search about a hotel booking found in a confirmation email. That’s intent no keyword would capture from the query alone. Under the new policy, the same inference that generates the organic answer can inform which ads appear.
But there’s a catch: you can’t buy this signal.
- No new targeting controls
- No audience segments tied to connected-app data
- No reporting to see whether a conversion came from this signal
- No published figures on how many users have opted in
The value — if any — flows through Google’s own ad selection. That makes measurement difficult and attribution murky.
The privacy angle for your brand
The change is opt-in, but it’s opt-in for a different feature. Users connect apps to Search or Gemini for better answers, not to enable ad targeting. Google’s policy describes no separate consent for the ad use.
Sensitive categories remain protected. Google still won’t use race, religion, sexual orientation or health inferences for personalized ads. But how it separates a medical appointment email from a hotel booking email isn’t explained.
This matters for your own privacy messaging. Users are trusting Google with deeply personal data in exchange for convenience. If your brand does the same, clarity about the trade is non-negotiable.
What to do about it
Don’t expect a new lever in Google Ads. Expect subtle shifts in auction behavior for opted-in users.
Test AI Mode placements now if you haven’t. Google confirmed ads are being tested there for connected-app users. Running small experiments before inventory scales gives you a read on conversion quality before competition arrives.
Monitor conversion quality, not raw volume. Inferred intent from inbox data may behave differently from classic search intent. A conversion from AI Mode isn’t necessarily the same as one from a keyword search.
Keep your first-party data sharp. If Google is now using first-party signals from its own ecosystem more aggressively, your own first-party data strategy matters more — not less.
The bottom line: Google’s ad engine just gained a new, invisible signal. It’s opt-in, unmeasurable and out of your control. Your job is to be ready when the auction shifts.
Source: PPC Land
