Google Revises Privacy Policy So Gmail, Drive and Photos Data Can Shape Ads via AI Features

Google has changed its privacy policy so that content from Gmail, Drive and Google Photos can help personalise the adverts a user sees, provided that user has connected those services to the company's artificial intelligence features.

The revision took effect on 16 October and was first identified by Android Authority, according to Decrypt. It qualifies a commitment Google has maintained for years: that material stored in its email, cloud storage and photo services would not be used to target advertising.

What the policy now says

The previous pledge was unconditional. Google said it did not show personalised ads based on content from Drive, Gmail or Photos.

The updated text keeps that sentence but attaches a condition. According to the revised wording reported by Decrypt, the exception now covers cases where a user opts to share data from those apps with other Google products that offer personalised advertising.

In effect, the promise still holds by default, but it no longer covers users who have linked those apps to other parts of the Google ecosystem that run advertising. The change was made without a prominent announcement, which is why it surfaced through a publication's review of the policy rather than through Google itself.

The role of Personal Intelligence

The trigger for the exception is Personal Intelligence, Google's effort to make its AI products more useful by letting them draw on a person's own data. Features such as AI Mode in Search and the Gemini assistant can, once connected, consult a user's email, documents and photos to produce more specific answers.

The practical appeal is straightforward. A user asking about an upcoming trip could, for example, get a response that pulls a hotel reservation out of their inbox rather than a generic list of results.

Under the new wording, the same connection that powers those answers can now also inform which adverts that person is shown. The data flows once the user opts in to the AI feature; the advertising use follows from that choice rather than from a separate, dedicated consent for ad targeting.

What Google says has not changed

Google has emphasised several limits on the new arrangement. Content from Gmail, Drive and Photos is still not used for ad targeting by default. Advertisers do not gain access to the contents of anyone's emails, files or pictures. And the company says it will not share personally identifying details without permission.

Taken together, those assurances mean Google is not broadly scanning private inboxes and photo libraries to sell advertising. The change is narrower: it affects people who have chosen to plug their personal data into Google's AI tools.

That distinction matters. Personalised advertising in this sense typically means using signals about a person's interests to decide which ads to display, rather than handing raw data to the businesses buying the ads. Google's position is that the advertiser sees an outcome, not the underlying content.

Why it matters

Even with those caveats, the revision marks a shift in how Google frames the relationship between its AI products and its advertising business. Users who adopt the company's AI conveniences may now be giving up a degree of ad privacy in exchange, whether or not they realise it at the time they connect their accounts.

The bundling is the central issue. Someone who links Gmail to Gemini in order to find a booking confirmation is making a decision about convenience. Under the revised policy, that same decision also widens the data available for ad personalisation. Privacy advocates have long argued that consent is most meaningful when each use of data is presented and agreed to separately, and this arrangement ties two uses together.

There is also the question of trust built over time. Google's commitment to keep Gmail, Drive and Photos content away from ad targeting was a longstanding reassurance, particularly for email, which can contain some of the most sensitive information people hold online. Adding a condition to that pledge, rather than retaining it in full, changes what users can assume about how their data is handled.

For Google, the commercial logic is visible. Advertising remains central to its business, and its AI products become more capable the more personal context they can access. Aligning the data used by AI features with the data available for ad personalisation brings those two priorities closer together.

Opting out, and what lingers

Users who want to avoid the new arrangement have options. They can disconnect Gmail, Drive or Photos under the Connected Content apps section of the Data & privacy settings in their Google Account. Alternatively, they can switch off personalised recommendations.

There is a complication, however. Android Authority noted that data collected before disconnecting may remain in a user's Search history even after the apps have been unlinked. That suggests disconnecting stops future use but may not fully erase what has already been gathered, and users seeking a clean break may need to review their Search history separately.

The main routes available are:

• Disconnecting the relevant apps under Connected Content apps in Data & privacy settings
• Turning off personalised recommendations
• Checking Search history for data that may persist after disconnection

A pattern of quiet changes

The update fits a broader pattern of scrutiny around how Google handles user data as it builds AI into its products. In May, Decrypt reported that Google's Chrome browser had begun installing a multi-gigabyte on-device AI model on users' computers without prominent notice. Separately, the company removed a privacy claim associated with its Gemini Nano model.

Both episodes drew criticism over transparency, and the latest policy revision shares a common thread with them: significant changes to how AI and personal data interact, introduced with little public signalling.

Google is not alone in racing to embed AI across its products, and the tension between making assistants more helpful and keeping personal data contained is an industry-wide challenge. AI tools that know more about a user can do more for that user, but the same depth of access raises questions about what else the data can be used for.

What to watch

Several questions remain open. One is how clearly Google communicates the advertising implication at the point where users connect Gmail, Drive or Photos to its AI features. If the link between AI convenience and ad personalisation is not made explicit during setup, criticism over informed consent is likely to grow.

Another is whether regulators take an interest. Changes to privacy commitments, especially those affecting email and personal photos, tend to attract attention from data protection authorities and consumer groups, though no such response was reported in the source material.

A third is user behaviour. If significant numbers of people disconnect their apps or switch off personalised recommendations in response, it could limit the value of Personal Intelligence to both users and Google. If few do, the change may become an accepted part of how AI assistants are funded.

Finally, the persistence of previously collected data in Search history is a detail worth monitoring. How Google handles that residual information will shape whether opting out is as effective as the settings suggest.