Google and Meta both started labeling AI-generated ads. Here's what changes

Two of the largest ad platforms on earth quietly rebuilt their disclosure systems inside the same month. On July 9, 2026, Google announced it would label ads made with generative AI across five of its products. Days earlier, Meta had already begun rolling out updated AI disclosure tags on Facebook and Instagram ads. Neither company framed this as optional. Both are racing to be ready before the EU AI Act's transparency rules take enforceable effect on August 2, 2026, and both are applying the labeling globally rather than region by region, because building two versions of an ad platform is more expensive than building one.

What Google's new AI ad labels actually do

According to Google's own announcement on its ads blog, the company introduced an "AI label setting" that rolled out gradually through July 2026 across Google Ads, Display and Video 360, Campaign Manager 360, Merchant Center and Ads Editor. Ads made using Google's own generative AI creative tools are labeled automatically. Advertisers who build creative with outside AI tools can label those ads manually through the same setting. On the consumer side, the disclosure surfaces in a "How this ad was made" panel inside My Ad Center, viewable on Search, YouTube and Discover, according to reporting from PPC Land and MediaPost covering the rollout. Google's stated rationale is regulatory: rules in the European Union, India and New York now require disclosure for ads containing certain AI-generated or AI-edited assets, and building one global labeling system is simpler than maintaining a patchwork of region-specific ones.

What Meta's disclosure tags do differently

Meta's approach, covered by Social Media Today and Marketing Dive, works through the existing "About this ad" element, accessible from the three-dot menu on any promoted Facebook or Instagram post. The label triggers automatically in two situations: when an advertiser uses Meta's own generative creative features, specifically Background Generation, Image Generation or Add Animation inside Ads Manager, or when Meta's detection systems identify C2PA metadata showing the asset was created or significantly edited with an external tool such as Photoshop's generative fill or Dall-E. One meaningful carve-out: if Meta's in-house AI tools do not result in a significant edit and the creative does not include a photorealistic human, no label is applied. Political, social issue and election advertisers face a stricter bar, and must disclose any digitally created or altered photorealistic image, video or realistic-sounding audio regardless of how minor the edit.

Why now: the EU AI Act deadline is doing the driving

The EU AI Act's transparency obligations for realistic AI-generated or AI-manipulated depictions of people, objects, places or events take enforceable effect on August 2, 2026, under guidance covered by legal trade press including Mondaq. Non-compliance penalties can reach up to 3 percent of a company's worldwide annual turnover, a number large enough that no platform of Google or Meta's size is treating it as a regional problem to patch later. That is also why both companies chose global rollout over an EU-only toggle: maintaining separate creative pipelines and ad review systems by region is more operationally expensive than applying one disclosure standard everywhere and adjusting strictness by jurisdiction where the law demands more, as with political ads.

Does this actually affect your ad performance?

Neither platform has published data showing a direct click-through or conversion penalty from the label itself, and it is early enough in the rollout that reliable third-party data does not exist yet either. The more realistic risk is not the label, it is what the label reveals. A product photo, a customer testimonial or a spokesperson video that reads as authentic but turns out to be AI-generated erodes trust in a way a small disclosure badge does not, particularly for direct-response and e-commerce creative where believability is the entire job of the asset. This sits alongside, not on top of, Google's separate August 17 bidding change to Target CPA and Target ROAS, which is an unrelated mechanic hitting the same accounts in the same month.

A compliance checklist for August 2026

1. Check the AI label setting in Google Ads and Ads Editor and confirm it reflects how your creative was actually produced, not the default.
2. Review any Meta ad using Background Generation, Image Generation or Add Animation, and confirm the "About this ad" tag is showing correctly.
3. If you run political, social issue or election ads on Meta, apply manual disclosure to any AI-altered photorealistic asset regardless of how small the edit, since the stricter carve-out has no minor-edit exception.
4. Flag any creative using an external AI tool for images or video and label it manually on Google, since automatic detection there is limited to Google's own tools.
5. Loop your content and creative team in now if they are producing AI-assisted assets for paid campaigns, so disclosure happens at creation rather than as an after-the-fact fix.

Frequently asked questions

What ads does Google now label as AI-generated?

Google automatically labels ads made with its own generative AI tools across Google Ads, Display and Video 360, Campaign Manager 360, Merchant Center and Ads Editor, visible in the "How this ad was made" panel inside My Ad Center on Search, YouTube and Discover. Advertisers can also manually label ads created with outside AI tools.

What triggers Meta's AI disclosure tag on Facebook and Instagram ads?

Meta applies the tag automatically when an ad uses Meta's own generative features such as Background Generation, Image Generation or Add Animation, or when its systems detect C2PA metadata showing the creative was made or edited with an external AI tool. The label appears in the "About this ad" section.

When does the EU AI Act require AI ad disclosure?

The EU AI Act's transparency obligations for realistic AI-generated or AI-manipulated content, including advertising, are enforced from August 2, 2026. Non-compliance penalties under the Act can reach up to 3 percent of a company's worldwide annual turnover, which is why Google and Meta rolled labeling out globally rather than only in the EU.

Will AI labels hurt my ad's click-through rate?

There is no public platform data yet showing a consistent CTR or conversion penalty from AI labels alone. The bigger risk is reputational: a label on a testimonial, product shot or spokesperson video that looks human-made but is not can damage trust more than the label itself.

Do I have to manually label ads I made with an external AI tool?

On Google, yes, for tools outside Google's own AI features, using the AI label setting in Ads Manager or Ads Editor. On Meta, manual labeling is recommended but the platform also runs automatic detection via C2PA metadata, so undisclosed AI content can still get flagged without a manual label.

What happens if I do not disclose AI-generated ad content?

Undisclosed AI content that is later detected can be labeled retroactively or restricted by the platform, particularly in regulated categories like politics, housing, credit and health. Under the EU AI Act specifically, non-disclosure exposes the advertiser to regulatory penalties, not just a platform-level warning.

The takeaway

This is not a creative-quality crackdown, it is a transparency mandate, and the two are easy to confuse. Google and Meta are not penalizing AI-made ads. They are making it visible which ones are AI-made, on a timeline set by regulators rather than by either platform's product roadmap. The advertisers who come out ahead here are the ones who audit their creative library now, understand exactly which assets will carry a label, and decide deliberately whether that label changes how the asset should be used, rather than finding out from a client or a compliance notice in September.