Google's Privacy Policy Update Opens Door to Ads Built on Personal AI Data

Connect your inbox to the AI, and the ads may follow
Google's updated policy lets Personal Intelligence users' connected data personalise ads, within stated limits.
Mark

So what actually changed at Google?

Mimi

The privacy policy now says that if you turn on Personal Intelligence, data from connected services like Gmail or Photos can be used to personalise content, including ads.

Mark

Does that mean advertisers can read my email?

Mimi

No. The policy says advertisers still can't access Drive, Gmail or Photos information, or PII like your name and email address.

Luke

Worth being careful there. That's what the policy states. The reporting doesn't explain how the separation works in practice, so we can't verify it from here.

Mark

And sensitive topics?

Mimi

Google says it will keep not showing ads based on race, sexual orientation, religion or health.

Luke

Again, a commitment, not an audit. Nothing in the piece tests it.

Mark

Why bring Meta into it?

Mimi

Meta tried to broaden its AI training terms this month and reversed quickly after backlash. It shows how sensitive this ground is.

Luke

But the two cases differ. Meta's concerned scraping sites; Google's concerns users who opt in. The article draws the contrast itself, noting Google already has crawl permission from most sites.

Mark

So what's unknown?

Luke

How many people will enable the feature, and how regulators will react. Neither is in the reporting.

  • Google's revised privacy policy now lets its search and Gemini tools draw on personal photos, Gmail and Drive data to tailor ads.
  • The change applies only to people who switch on Personal Intelligence, but the opt-in line is already drawing scrutiny.
  • Google insists advertisers will never see raw Drive, Gmail or Photos content, nor names or email addresses, and sensitive topics stay off-limits.
  • The timing sharpens the unease, following Meta's briefly published AI training terms that it reversed after a public backlash.
  • Regulators and users are now watching closely to see whether AI-driven personalised advertising becomes accepted practice or triggers a fresh push for limits.

Google has quietly rewritten the terms that govern how its services share information with one another, opening a path for the personal material people store in their photos, inbox and cloud files to shape the advertisements they see. The shift arrives in a period when the question of who owns personal data, and who may profit from it, is being argued across the technology world, often after the fact and under public pressure. Google presents the change as a convenience for users who choose to connect their assistant to their own records, while keeping certain boundaries in place for advertisers and sensitive subjects. Whether that bargain between helpfulness and surveillance holds will depend on how users and privacy regulators choose to read it.

Google has changed its privacy policy so that people who turn on a feature called Personal Intelligence can be shown ads based on their personal data. The change was first noticed by Android Authority, and it marks a notable shift in how the company says it will use information people share with its AI tools.

Personal Intelligence allows users to connect Google's AI assistants to their email and photo libraries, so the assistant can offer more useful answers and recommendations. The updated policy now makes clear that this connected information can also inform advertising. Search or the Gemini apps may reach into Photos, Gmail and Drive when a user allows it, and Google's example is simple: a search for mountain bikes could lead to sports equipment ads on YouTube.

Google says some protections remain. Advertisers will not gain access to Drive, Gmail or Photos content, or to identifiers such as names and email addresses, and ads will continue to avoid sensitive subjects like race, sexual orientation, religion and health.

The move comes at a sensitive time. Technology companies have faced criticism for training AI systems on material gathered from across the internet, and some have revised their terms to use customer data in similar ways. Meta recently drew a strong reaction after a revised terms page appeared to permit broad scraping for AI training, and it quickly reversed the change. Observers note that Google's position may differ somewhat, since websites generally expect Google to crawl them for search. What happens next will depend on how users and privacy regulators respond to ads built on personal data that people have chosen to connect.

Google has changed its privacy policy so that people who switch on a feature called Personal Intelligence can be shown targeted ads based on their personal data. The change was first spotted by Android Authority.

Personal Intelligence lets users connect Google's AI tools to their email inbox or photo library. The idea is that the AI, knowing more about the person, can give better responses and recommendations. The updated policy now spells out that this connected data can also feed into advertising.

The policy language is plain about it. Users may allow certain Google services, such as Search or the Gemini apps, to reach data held in other Google services, including Google Photos and Workspace apps like Gmail and Drive. That access is meant to provide a more personalised experience, and it is subject to the policies and settings of the service doing the accessing. As an example, the policy says Search may process some of the data a user allows it to see in order to personalise content, including ads.

A second passage covers ads more broadly. Depending on their settings, users may see personalised ads based on their interests, activity, or personalised content from Google services. Google's own illustration: someone who searches for mountain bikes in Search might then see ads for sports equipment on YouTube.

There are limits. Advertisers still will not be able to access a user's Drive, Gmail and Google Photos information, nor other personally identifiable information such as a name or email address. Google also says it will continue to avoid showing ads based on sensitive topics involving race, sexual orientation, religion or health.

The update lands at a tense moment for data ownership. Tech companies have been scraping publicly available, copyright-protected material from across the internet to train AI models. Some have also quietly revised privacy policies and terms of service so they can use customer data for training, or license it to other companies for the same purpose.

Meta offered a recent example. Earlier this month it tried to revise its terms in a way that appeared to allow broader AI training, and the backlash was considerable enough that the Facebook and Instagram parent quickly reversed course. Its commercial terms of use page had been updated with language reportedly suggesting it could scrape content from sites, news sites among them, that have a Meta business account, to train its AI models. Legal experts, including people at Neiman Lab and James Grimmelmann, a Cornell professor of digital and information law, noted last week that the now-deleted provision would have let Meta use scraped content for free and override the blocking instructions in a site's robots.txt file.

The report notes that such a move may be harder in Google's case, because websites broadly allow Google to crawl their content for indexing in Search. What to watch next is how users and privacy regulators respond to AI-driven ad personalisation built on data people have chosen to connect.

Search may process some of the data that you allow Search to access to personalize content, including ads.
— Google's updated privacy policy
Contáctanos FAQ