Gmail data can now shape your ads

Google’s privacy policy — the one at policies.google.com/privacy, effective October 16, 2026 — has quietly picked up a carve-out. The sentence that has stood between Gmail, Drive, and Google Photos content and the ad system since July 2021, “We don’t show you personalized ads based on your content from Drive, Gmail, or Google Photos,” is still there. It now ends with an “unless”: unless you choose to make data from these services available to other Google services that support personalized ads. The switch that triggers it is Personal Intelligence, the opt-in data connection behind Gemini apps and AI Mode search. Android Authority broke the news on October 23; The Verge, Ars Technica, and PPC Land each followed the same day. Chinese-language coverage has not appeared yet, which makes this a first-pass analysis in that market.

Key points

The pledge stayed; it gained an exception. The policy now reads “…unless you choose to make data from these services available to other Google services that support personalized ads,” effective October 16, 2026.
The indirect path is in writing: Google says ads “will not be based directly on your Gmail inbox or your Google Photos library,” while Search may use “summaries, excerpts, generated media, and inferences” to serve them.
There is an off switch: disable Personal Intelligence in the Gemini app’s settings, or disconnect Gmail, Calendar, and Photos under your account’s connected content apps.

What changed in the text

Comparing against the policy archive shows the full extent of the edit. PPC Land diffed the October 16 text against 54 earlier versions of the policy, an archive reaching back to June 1999. The pledge first appeared on July 1, 2021 and survived 17 versions untouched; the previous version took effect October 1, 2026, just fifteen days before this one.

There are two changes, not one. The second is subtler: the policy’s grounds for ad personalization grew from two bases to three. Alongside your interests and your activity there is now “personalized content from Google services.” A footnote on the policy page also states that Search “may process some of the data that you allow Search to access to personalize content, including ads.” Grant AI Mode access to your inbox, and that same grant covers ad personalization.

What others say

The outlets read the edit differently, and the disagreement is the interesting part. Ars Technica (Ryan Whitwam, October 23) frames it as a dismantled firewall: an explicit prohibition replaced with equivocal wording that lets piped-in data inform targeting. The Verge (Jess Weatherbed, October 23) and the policy text itself say the pledge survived and merely gained an exception. We checked the live policy page: the sentence is intact, so Ars’s “removed” framing does not match the primary source — we side with the text.

Google’s spokesperson response lands between the two: connecting apps is “entirely opt-in,” ads “will not be based directly on your Gmail inbox or Photos library,” and nothing goes to third-party advertisers. But the Google help page PPC Land quotes is more specific: Search works from “summaries, excerpts, generated media, and inferences from your relevant media, emails, and files,” and “that same info may be used to show you ads.” A short email’s summary, the page concedes, may just be the email. The policy still excludes sensitive categories — race, sexual orientation, religion, health — from ad targeting; per Google’s account as relayed by The Verge, advertisers get no access to Drive, Gmail, or Photos content or identifying details.

Timing matters too. PPC Land had already reported, on October 20, Google’s confirmation that ads were being tested in AI Mode for Personal Intelligence users from around October 16. The policy edit and the ad test landed the same day.

How to turn it off

The off switch lives in Google’s own documentation. To stop the data connection, disable Personal Intelligence in the Gemini app’s settings, or go to Data and privacy, then Personalization settings, then Manage personalized recommendations, then Connected content apps, and disconnect Workspace and Photos. One caveat worth stating plainly: disconnecting is not deleting. PPC Land notes that data already collected stays, and can still be used for product improvement and AI training.

Our take

For anyone building on Google’s AI stack — and for anyone with a decade of mail in Gmail — the assumption to revisit is the data flow, not the wording. Every earlier guarantee that inbox content stays out of the ad system now holds only while Personal Intelligence is off. Meta removed a “training” clause from its commercial terms before they took effect (our earlier coverage); Google went the other way, keeping the clause and bolting on an exception. Different edits, same outcome: the default defense yields to optional consent. Regulators have a live benchmark to compare against — Norway just fined Microsoft’s Xandr over sensitive data leaking into real-time bidding (our earlier coverage), and Personal Intelligence itself expanded to nearly 200 countries after Google I/O in May. Whether this exception draws the same regulatory attention is the thing to watch. Until then, the one lever users hold is the toggle: leave it off, or switch it back off.