Google Ads now requires disclosure labels on AI-generated ad creative

Google Ads is rolling out mandatory disclosure labels showing whether generative AI was used to create or edit an ad's creative. If the asset came out of Google's own generative AI tools, the label is applied automatically. If it came out of a third-party tool, the advertiser has to declare that through a new control. Search Engine Journal reported the rollout on 13 July 2026.

What the label covers

The disclosure spans most of Google's advertising surface: Search, YouTube, Discover, Display & Video 360, Campaign Manager 360, Merchant Center and Ads Editor. It arrives gradually through July 2026.

On the user's side, the detail lives in My Ad Center, in a panel called "How this ad was made".

Automatic for Google's tools, manual for everything else

This split is the whole story operationally.

Google knows when its own generative tools touched an asset, so it labels those without anyone asking. It has no way of knowing that a designer generated the background in a third-party tool, or extended an image to fit a placement, or cleaned up a product shot with a generative fill. That declaration is the advertiser's job, through the new control.

Which means the label's accuracy depends entirely on whether the person uploading the ad knows what happened to the file three steps upstream. That is not usually a documented thing.

Where the labels appear

Labels display to users in jurisdictions that have a legal requirement for them. Google named the EU, India and New York State. The announcement says "including", so that isn't necessarily the complete list, and Google didn't publish one. It also said nothing about Thailand in either direction, so treat the named three as the floor rather than the ceiling.

The part that lands on the agency

Google's control gives advertisers a switch. It doesn't give them the answer to when the switch should be on.

Deciding whether a given piece of creative counts as AI-created or AI-edited, and whether that triggers a disclosure requirement in a given market, is left to the advertiser and whoever runs the account. A generative fill that removed a power socket from a product photo, an upscaled hero image, a background replaced in a third-party tool: someone has to make a call on each of those, and Google isn't making it.

That is a documentation problem before it is a legal one. Nobody can declare what nobody recorded.

What this means for Thai marketers

Two situations, two levels of urgency.

If your campaigns only ever serve inside Thailand, the label may never render for your audience. That is not a reason to skip the declaration. You don't control where Google's requirement list goes next, and creative built for one market gets reused in another far more often than any media plan admits.

If you run anything that reaches the EU, India or New York State, and plenty of Thai advertisers do (export brands, hotels selling to European travellers, anyone with an international media plan), the declaration isn't optional and the label will be visible on your ads.

The awkward case is the in-between one: a Thai brand whose agency, freelancers and production partners all use generative tools casually, on a campaign that might get extended internationally next quarter. That's where the paperwork gap bites.

What to change in the workflow

• Ask the AI question at brief stage, not at upload stage. By the time an asset is in Ads Editor, the person clicking upload usually has no idea how the image was made.
• Record it per asset, not per campaign. One AI-extended background in a set of twelve images is still an AI-edited asset.
• Get it in writing from freelancers and production houses. A verbal "we just touched it up a bit" isn't a record you can act on.
• Cover Merchant Center and Ads Editor too. The disclosure isn't only a Google Ads UI thing, and product imagery is exactly where generative editing sneaks in quietly.

FAQ

Do we have to label ads made with Google's own AI tools?

No. Those are labelled automatically. The manual control exists for third-party tools.

Does this apply to creative we only edited slightly with AI?

Google's wording covers creative that generative AI was used to create or edit. The announcement doesn't carve out light edits, so don't read it narrower than Google wrote it.

What happens if we don't disclose?

Google didn't publish a penalty in this announcement. The exposure is legal rather than platform-side: the named jurisdictions have the requirement, and the advertiser is the party making the call.

How does a user see the detail?

By opening the "How this ad was made" panel in My Ad Center.

The short version

This isn't a big change to how ads are built. It's a change to what has to be known about how they were built, and by whom, before launch. Accounts that already track creative provenance will barely notice. Accounts where "the designer made it" is the whole audit trail have some work to do. If you want your Google Ads creative approval process checked against this before the rollout finishes, that's a short conversation. Our Google Ads team can walk through your current workflow and flag where the AI declaration would fall through.