Google is adding a label to the ads its AI helps create. In a July announcement, the company said a new “How this ad was made” panel in My Ad Center will indicate when generative AI was used to create or alter an ad, accessible from the three-dot menu or info icon on ads across Search, YouTube and Discover. When advertisers use Google’s own generative tools, the disclosure is added automatically. When ads are made elsewhere, a new control lets advertisers flag AI involvement themselves, and depending on local requirements a label may appear directly on the ad.
The move builds on plumbing Google already had in place. The company embeds imperceptible SynthID signals in outputs from its generative AI tools, and since 2023 it has required disclosure of synthetic or digitally altered content in election advertising. Existing policies already prohibited misleading or deceptive ads whether made with AI or not; the new panel is about provenance rather than quality — telling a viewer not whether an ad is good, but what made it.
The timing is not accidental. Google’s September Demand Gen updates pushed AI further into the buying process, adding conversational Business Agent experiences alongside YouTube video ads, one-tap landing page visits from image ads on Shorts and Gmail, and affiliate location extensions in Google Maps. Advertisers adding Gmail to Demand Gen campaigns saw image creatives deliver a reported 40 percent increase in conversions at the same return on investment. In other words, as AI generates more of the creative and more of the conversation around it, disclosure becomes part of the product rather than an afterthought.
For advertisers, three implications stand out. First, the disclosure is automatic only inside Google’s own tools; creative produced in third-party systems needs a human to declare it, which makes AI-use documentation part of trafficking, not a legal footnote. Second, provenance metadata like SynthID means platforms may be able to identify AI-generated assets even when nobody declares them — an audit trail brands should assume exists. Third, labels change the context in which creative is judged. An ad a viewer knows was AI-assisted will be read differently, and categories built on trust — health, finance, local services — should expect regulators and platforms alike to demand clarity.
None of this slows the automation itself. Google is not asking permission to use AI in advertising; it is standardizing how everyone finds out. For marketing teams, the action item is procedural: inventory where generative AI touches your creative, decide who declares it, and make sure the declaration survives the handoff from studio to platform. The label is coming either way. The only question is whether your team applied it or the platform did.
Related reading: Connected TV Takes the Growth Crown as Buyers Move Budgets Out of Linear · TikTok Shop Is Becoming a Storefront, Not Just a Feed · Retail Media's $71 Billion Year Runs Through Two Checkouts: Amazon and Walmart
1 Comment