Google is extending AI disclosure beyond election ads to the whole marketplace. Starting Thursday, ads served on Search, Discover, and YouTube will carry a “created or edited with AI” note when Google’s own generative tools were used to make them. You’ll find it under the “how this ad was made” tab inside “My Ad Center.”
On paper, this is the sensible next step. Google already bans misleading and deceptive ads, and it already forced political advertisers to disclose synthetic content. Broadening that requirement to every ad acknowledges what everyone in the industry already knows: generative tools are now baked into ordinary commercial creative, not just campaign attack spots.
The disclosure nobody will see
The catch is placement. This label doesn’t sit on the ad. It sits behind a two-click journey into My Ad Center — a settings surface the overwhelming majority of users have never opened and will never think to open. A transparency feature that requires you to already suspect an ad, then go hunting through account menus to confirm it, is not really designed to change what a casual viewer believes in the moment. It’s designed to be available.
That distinction matters. The harm from synthetic ad content — a too-perfect product shot, a fabricated demo, an AI-smoothed “before and after” — lands at the point of impression, in the scroll, before anyone questions anything. A disclosure gated behind a menu does nothing at that moment.
Scope Google conveniently controls
There’s a second limit worth naming. The reporting indicates the label triggers automatically when Google’s AI tools are used. That’s clean to detect — Google owns the pipeline — but it quietly defines the problem down to the slice Google can see. An advertiser who generates a synthetic image in a third-party tool and uploads the finished file is far harder to catch, and it’s not obvious this system catches them at all. So the ads most likely to be deceptive — made by someone deliberately routing around scrutiny — may be exactly the ones that stay unlabeled.
Still, a floor worth having
None of this makes the move worthless. Establishing that “AI-made” is a disclosable property of an ad, industry-wide, is a foundation you can build on. Labels can migrate onto the ad surface itself. Detection can extend to uploaded creative. Regulators now have a baseline to point at. Google has planted a flag where before there was nothing but a policy PDF.
But let’s be precise about what shipped: a compliance-grade audit trail, not a consumer-facing warning. It satisfies the people asking “does Google disclose this?” far better than it protects the person actually looking at the ad. The gap between those two is where the next round of pressure — from users, journalists, and regulators — should land.