Google Will Label Which Ads Were Made With AI

Google has started showing users whether an ad they're looking at was made using AI. The feature is built into the "My Ad Center" panel and is rolling out to Search, YouTube and Google Discover globally, the company said. It's the first attempt to bring the average user information that the platform had, until now, only required to be disclosed in one very specific type of advertising.

What exactly changes

Until now, Google only required disclosure of AI use in election ads. The rest of its advertising could rely on synthetic or digitally altered content without consumers ever knowing. The new measure extends that transparency to any ad category.

The mechanism is simple. The "My Ad Center" panel already let users block or report ads, learn about the advertiser, or understand why a particular ad was shown. Now, by clicking the three-dot menu or the info icon that accompanies ads, users will see a new option: "How this ad was made," which indicates whether the ad was created or edited with AI.

Google frames the decision around a specific problem. AI makes it easier for businesses to create ads, place their products in various settings, and save money on real-world e-commerce photography. But, as the company itself acknowledges, that can be misleading if consumers don't know that what they're looking at isn't a real product photo. The platform prohibits misleading and deceptive ads, but that rule didn't stop an ad from using synthetic images without disclosing it.

The weak point: Google doesn't verify anything

The detail that shapes the entire measure lies in how the label gets triggered. When an advertiser uses Google's generative AI advertising tools, the disclosure will be automatically enabled. The platform knows the content came from its own systems and declares it without human intervention.

The problem arises when the ad is created outside Google's ecosystem. In that case, the advertiser needs to use a new control to indicate if AI was involved in its creation. And here's the key: Google will not perform its own check to determine if that's the case. Transparency, then, depends on the good faith of whoever is paying for the ad.

This turns a measure that sounds robust into something more fragile than it appears. An advertiser that generates synthetic images with any tool outside Google's and decides not to disclose it will face no pushback from the platform. The label will work fine for content "native" to Google's tools, but it leaves an obvious gap for everything else.

The company adds that, in some markets, the ad may also be labeled as AI-generated if local law requires it. In other words, wherever regulation makes it mandatory, the label stops being voluntary.

Why this move matters

Advertising is one of the areas where generative AI has moved in fastest and with the least noise. Producing a convincing product image now costs a fraction of what a photo shoot used to cost, and the result can place an item in any imaginable setting without a camera or a studio. That savings is real and explains much of the adoption.

The flip side is consumer trust. When someone sees a product photo in an ad, they implicitly assume it represents something real. If that image is synthetic — an object that might not actually look that way, set in an environment that doesn't exist — the line between legitimate advertising and deception blurs. Google's label attempts to give users back the ability to gauge what they're seeing.

There's also a defensive angle at play. Tech platforms have been under regulatory pressure over the spread of synthetic content for a while now, and getting ahead of it with a voluntary label is a way to stake out ground before regulation forces the issue. Google's own mention of markets where the law requires it acknowledges that regulation is already ahead in some places.

The limits of a label the advertiser controls

The value of a disclosure depends on three things: whether it's visible, whether it's reliable, and whether users actually look for it. Google's measure checks the first box in a low-key way — you have to open a menu to see it, it doesn't appear over the image itself — and partially fails the second, since it delegates truthfulness to the advertiser without verification.

That puts it far behind more ambitious systems like technical content provenance marking, where information about AI use travels embedded in the file itself and can't simply be removed. Google's solution is more like an honesty checkbox: useful, but only as reliable as whoever checks it.

It also remains to be seen how many users will actually click the three-dot menu to find out how an ad was made. Most people don't interact with those controls. A label tucked behind two clicks informs those who already suspect something, but it's unlikely to change the experience of users who simply scroll through their feed.

What to watch going forward

The big question is whether Google will eventually tighten the system to verify content created outside its own tools, or whether the label will keep depending on self-disclosure. The difference between those two paths is the difference between real transparency and a mere image gesture.

The other front is regulatory. By explicitly mentioning markets where the law requires labeling, Google acknowledges that its voluntary measure coexists with legal frameworks advancing at different speeds depending on the country. As more jurisdictions require disclosure of synthetic content in advertising, the voluntary part of the system will keep shrinking until it becomes mandatory by default.

For advertisers, the immediate message is that disclosing AI use is shifting from a strategic choice to, gradually, a market expectation. For users, the tool exists — but whether they know it's there and use it will be up to each one of them.