GOOGLE ADDS AIGC LABELS ACROSS SEARCH AND YOUTUBE, BUT THERE'S A CATCH

Google is rolling out a transparency feature that tells you when an ad you are seeing was created or edited with generative AI. The disclosure appears in the My Ad Centre panel under a new section called accessible by tapping the three dots or info button on a Google Search, YouTube, or Discover ad.

There is an important asymmetry baked into the rollout. If the advertiser used Google's own generative AI tools, the label gets applied automatically. If they used someone else's tools, the disclosure is left to their discretion: an opt-in toggle that Google says is available for brands that want to use it.

you may like

GOOGLE IS NOW TRAINING ITS AI ON YOUR UPLOADS, YOU'RE OPTED IN ALREADY

GOOGLE ADDS YOUTUBE MUSIC, INSTACART & CANVA TO AI MODE SEARCH

GEMINI LETS ANYONE* GENERATE PERSONALISED IMAGES FOR FREE

The company framed the move as helping people navigate ads while giving advertisers clear tools. "We want to help people better understand the ads they see, while providing advertisers with straightforward tools to navigate evolving industry standards"

WHAT GETS LABELLED

The label appears in the section of My Ad Center, which is already where Google shows info about why you are seeing a particular ad and who placed it. Users can reach it by clicking the three-dot menu or the info icon on ads across Search, YouTube, and Google Discover.

For ads created with Google's own generative AI tools, such as those that generate product images or ad copy, Google handles the disclosure server-side, so the advertiser does not need to do anything extra. For ads made with third-party AI tools, the advertiser can toggle a setting to indicate that generative AI was used. Google does not require it.

In some regions, local regulations may push that label directly onto the ad itself, not just inside the panel. Google's language says the label "may also appear directly on the ad" depending on the jurisdiction and whether the advertiser opted in or the requirement is mandatory.

WHAT CHANGED AND WHAT DID NOT

Until this update, Google only required AI disclosures for election ads, a rule introduced in 2024 for synthetic or digitally altered content that could mislead voters. Everything else -product ads, brand campaigns, service promotions- had no transparency requirement at all.

That gap matters because AI tools have made it trivial to generate photorealistic images of products in settings that never existed. A furniture brand can render a sofa in a lavish living room without renting a studio. A clothing retailer can put a model in a cityscape that was never photographed. These ads save businesses money and speed up production, but they also leave buyers guessing whether the image they are looking at is real.

Google's own advertising policies already prohibit misleading and deceptive ads, so a fake product photo that materially misrepresents what a buyer gets would still be against the rules. But an ad can be AI-generated without being deceptive. It might simply show a product in a generic scene, and that is where the line gets fuzzy. The label is meant to clear it up.

HOW IT COMPARES TO META

Meta already ships a similar label inside its panel across Facebook and Instagram. The core idea is the same: let users see whether AI was involved in making an ad, rather than banning AI from ads outright. For Google, the label is automatic only for its own tools and opt-in for third-party tools, which means the system relies on advertiser honesty.

WHAT THIS MEANS FOR ADVERTISERS AND USERS

For advertisers who use Google's AI tools, the change is frictionless — disclosure happens in the background. For everyone else, the opt-in toggle adds a layer of voluntary compliance. Google is essentially betting that enough brands will choose transparency to make the system credible, rather than requiring it.

The real question is whether the default human impulse is to toggle disclosure on. In marketing departments under pressure to produce volume, the path of least resistance is often to leave the toggle off and let the ad run without the label, especially if the label might introduce consumer scepticism. Google will need to show that the system does not punish labelled ads with lower performance or higher costs if it wants disclosure rates to be meaningful.

THE PUSH FOR TRANSPARENCY

Google earlier this year expanded access to SynthID and C2PA content credentials, digital watermarking tools that can help identify deepfakes and AI-generated media. Those tools operate at the file level, embedding metadata that persists across platforms. The ad label is a separate layer, aimed at the moment of consumption rather than the moment of creation.

Together, they form a piecemeal approach: SynthID and C2PA address the supply chain, while My Centreter addresses the user interface. Neither is comprehensive on its own, but the combination covers more ground than any single fix could.

CAN IT BE TRULY IMPLEMENTED

The opt-in nature of third-party disclosure is the vulnerability. For the system to work at scale, it needs adoption beyond Google's own tools. The next logical step would be to make the AI disclosure mandatory for all ads, regardless of tool origin, the way Google already handles political ads. That would bring enforcement teeth and close the loophole.

what to read next

META'S MUSE IMAGE DEFAULTS TO USING YOUR PUBLIC INSTAGRAM FOR AIGC

SWEDISH COURT ORDERS GOOGLE TO PAY KLARNA $1.46BN IN ANTITRUST DAMAGES

NOTEBOOKLM NOW GEMINI NOTEBOOK, ADDS CLOUD CODE EXECUTION

For now, if you see an ad on Google Search and want to know whether AI made it, you can check the three-dot menu. If the label is missing, the ad either used no AI or the advertiser chose not to tell you.