On July 9, 2026, Google added a new panel to My Ad Center that 2.4 billion users can now access by tapping the three-dot menu on any Search ad. The panel is labeled “How this ad was made.” It tells users whether the ad was created or edited with generative AI. The rollout spans Search, YouTube, and Discover, and it’s not optional for advertisers using Google’s own AI creative tools. If you use AI Max for Search, Demand Gen image generation, or asset generation in Performance Max, the disclosure appears automatically. If you use third-party AI tools like Midjourney or ChatGPT for creative, Google has added a self-declaration control that puts the responsibility on you [1].
AI creative tools have gone from experimental to default across European Google Ads accounts in roughly eighteen months. This announcement matters because it is the first time Google has built a consumer-facing transparency layer that connects directly to the EU AI Act, which takes full effect on August 2, 2026.
Table of Contents
What the “How This Ad Was Made” Disclosure Does
Keerat Sharma, Google’s VP and GM for Ads Privacy and Safety, published the announcement on July 9. Google is adding a new section to the My Ad Center panel, accessible globally through the three-dot menu or info icon on any ad across Search, YouTube, and Discover [1]. The panel indicates whether the ad was created or edited with AI.
This is the same panel that already shows why a user saw an ad and who paid for it. What is new is the production layer: not just who is targeting you, but how the creative in front of you was made. There are two paths for how an ad gets flagged.
Two Disclosure Paths: Automatic vs Manual
- When advertisers use Google’s generative AI advertising tools, the disclosure is added automatically. This covers AI Max, asset generation in Performance Max, Demand Gen image creation, and the animated image features announced at Google Marketing Live 2026.
- When advertisers create ads using AI tools outside of Google’s ecosystem, they must use a new control to declare that AI was used. Google is not detecting this automatically. The advertiser has to self-report.
The distinction matters. If your creative team uses ChatGPT to write headlines or Midjourney to generate display assets, you are responsible for declaring that in Google Ads. Google won’t catch it for you.
EU AI Act Deadline: August 2, 2026
The My Ad Center panel is global. Every user in every market can access it. But there’s a second layer. In some jurisdictions, a label may appear directly on the ad itself. Google has named three specific regions where on-ad labels are required by local law: the European Union, India, and New York State [2].
For European advertisers, this is where the EU AI Act connection becomes concrete. Article 50 of the AI Act requires providers of AI systems to ensure transparency when AI-generated content is disseminated. The regulation takes full effect on August 2, 2026. Google’s on-ad labeling is the compliance mechanism that helps advertisers meet that obligation. If you’re running AI-generated ads in France, Germany, Sweden, Denmark, Finland, or Ireland, expect to see labels on those ads starting this summer.
Norway and the UK are not on the named list for on-ad labels. Norway is EEA but not EU, and the UK has its own regulatory approach to AI. The My Ad Center disclosure still applies in those markets, though. Users there will still see the transparency panel if they look for it. Norwegian advertisers running campaigns into EU markets are covered by the EU rules for those campaigns regardless of where the account sits, which makes treating the whole EEA as in-scope the safer operating assumption.
What Advertisers Must Do
Google’s position is clear: advertisers “are ultimately responsible for determining when AI use requires a disclosure” and for compliance with local law [2]. You can’t wait for Google to flag your AI-generated creative. You need an internal process. Four steps cover the ground.
- Audit your creative pipeline. Which assets are generated or edited with AI? Text, images, video, audio, all of it. If ChatGPT wrote your responsive search ad headlines, that’s AI use. If your designer used Adobe Firefly to generate product backgrounds, that’s AI use too.
- Map AI tools to disclosure requirements. Google’s own tools auto-disclose. Third-party tools require self-declaration. Document which tools your team uses and where they fall in Google’s framework.
- Update your ad review workflow. Before campaigns go live, someone should confirm whether AI was used and whether the appropriate disclosure has been made. Add it to your QA checklist.
- Prepare for on-ad labels in EU markets. Google will apply labels automatically or via your self-declaration. Make sure your creative team knows these labels will be visible to users.
The timeline is tight. The EU AI Act takes effect August 2. Google’s feature is rolling out gradually throughout July across Google Ads, Display & Video 360, Campaign Manager 360, Merchant Center, and Ads Editor [2]. Merchant Center being on that list is easy to overlook: it pulls feed and CSS operations into scope, so retailers generating or rewriting product data with AI have the same obligation as anyone building display creative.
Self-Declaration Is a Transitional State
The self-declaration control is the part of this rollout most likely to be misread as permanent. Right now, third-party AI use runs on an honor system: Google asks, the advertiser answers, and nothing checks the answer. Treating that as a stable arrangement is a mistake, because Google has already moved past it on another surface.
In May 2026, YouTube began automatically detecting undisclosed photorealistic AI content and labeling it for viewers, while also moving existing AI disclosure labels into more prominent positions [5]. That is the same company, the same provenance problem, and a decisive shift from asking creators to declare AI use to detecting it whether they declare it or not.
The infrastructure to do the same in ads already exists. SynthID embeds imperceptible watermarks in outputs from Google’s own generative models, and the detection side of that system is what makes automatic labeling possible on YouTube. Third-party tools are harder, but not indefinitely: photorealistic AI imagery carries statistical signatures, and detection models improve while old creative stays in the account.
The practical consequence is about sequencing. An undeclared AI asset is low-risk today and higher-risk the moment detection catches up, and it is the same asset either way. Campaigns built now with no record of what was generated leave an account with a backlog nobody can reconstruct later. Advertisers who log AI use from the start are not doing extra compliance work; they are avoiding an audit that has to be done from memory.
What a Disclosure Record Should Contain
“Document your AI use” is the advice everyone gives and almost nobody specifies. The record only has to answer one question reliably: for any asset currently serving, what was AI’s involvement and where was that declared? Six fields cover it.
- The asset as Google sees it. The asset ID or filename as it appears in Google Ads, Merchant Center, or Ads Editor — not the internal design-file name. If the record can’t be joined back to the serving asset, it can’t be checked.
- The specific tool and version. “Midjourney v7” or “Gemini via Asset Studio,” not “AI.” Tools change behaviour between versions, and a regulator’s question is about a particular output, not a category.
- The kind of involvement. Generated from scratch, edited, upscaled, background-replaced, or text rewritten. Both Article 50 and Google’s own control distinguish content that was created with AI from content that was edited with it, so a record that collapses the two loses the distinction the rules turn on.
- Who produced it and when. Obligations start on fixed dates. An asset generated in June 2026 and an asset generated in September 2026 sit on opposite sides of the AI Act deadline.
- Where the disclosure lives. Auto-disclosed by Google, self-declared on a given date, or assessed as out of scope with the reason recorded. “Out of scope” is a legitimate answer; an unrecorded assumption is not.
- Which markets the campaign serves. This determines whether an on-ad label is expected in addition to the My Ad Center panel.
A spreadsheet is enough for most accounts. Larger operations will want this as a field on the asset record in whatever system already holds creative, because the value comes from it being queryable, not from it being thorough. The failure mode worth designing against is not a missing tool but a record that lives in one person’s head. The record that matters is the one that can be produced without the person who made the asset.
How This Fits Google’s Broader Transparency Push
This isn’t Google’s first move on AI transparency in ads. In 2023, they introduced a requirement for election advertisers to disclose synthetic content. They also built SynthID, a watermarking technology that embeds imperceptible signals in AI-generated outputs. The “How this ad was made” disclosure builds on that foundation [1].
The pattern is consistent. Google is positioning itself as the platform that makes AI use visible to users. This matters for trust, and the same theme runs through how AI Mode and AI Overviews work in Search. Users want to know when AI is involved, whether it’s the search results they’re reading or the ads they’re clicking. Transparency builds confidence.
Google is also reducing its own regulatory risk. By building disclosure mechanisms now, they’re showing regulators they take the EU AI Act seriously. If Google can demonstrate that it’s making AI use clear to users, it becomes harder for regulators to argue for stricter intervention.
The arc is easier to read as three layers added over three years. In March 2023, the Ads Transparency Center made every ad from a verified advertiser across Search, YouTube, and Display publicly browsable, which answered who is advertising [6]. Advertiser verification answered who they are. “How this ad was made” answers how the creative was produced. Each layer took information that had been available only to the advertiser and made it available to anyone who looks.
The same direction shows up in the ad labels themselves. Google has been testing a shaded light-gray container behind the “Sponsored results” heading, making paid listings more visually distinct rather than less [7]. That reverses years of gradually softening ad labels, and it matches Bing’s move back to a bolder Sponsored label. For advertisers the read-across is straightforward: assume anything that distinguishes an ad from organic content, including an AI label, becomes more visible over time rather than quietly disappearing.
Impact on Your PPC Strategy
The disclosure doesn’t change the mechanics of building a PPC operating system with AI agents, but it adds a governance layer. If those agents are generating ad creative, someone needs to track which outputs are AI-generated and ensure they’re properly disclosed.
This is where the operational side of AI in advertising gets real. It’s not just about generating more variants or testing faster. You need to know what you used AI for and be able to prove it when asked. For large accounts with hundreds of campaigns, that means building documentation into your workflow from day one.
The disclosure also has implications for creative testing. If users can see that an ad was AI-generated, does that affect click-through rates? There is no data yet. Google hasn’t published any research on user response to the “How this ad was made” panel. It’s worth monitoring. If AI-labeled ads underperform, you may need to adjust your creative strategy or reconsider how much AI you’re using in production.
The Terms of Service Change Behind It
The disclosure panel arrived a month after a quieter change that gives it contractual teeth. On June 3, 2026, Google updated the Google Ads Terms of Service for the first time in eight years, largely to address AI [4]. The update is what converts “you should disclose” into “you agreed to be responsible.”
Two clauses matter for AI creative. The terms now clarify how advertiser-supplied inputs can be used to improve campaign performance, and they restate that the advertiser holds responsibility for rights relating to both those inputs and automated campaign assets [4]. That second point is easy to skim past and expensive to miss: responsibility for rights covers assets Google’s own tools generated. Automatic disclosure in My Ad Center says the asset was made with AI. It does not say the asset was cleared for use, and it does not move the rights question onto Google.
The same update adjusted liability caps, introduced new fee disclosures, allowed batch arbitration for similar claims, and removed the fallback support email [4]. Read together, the direction is consistent: Google is building the tooling that makes AI use visible while tightening the contractual language that keeps consequences with the advertiser.
This changes who inside an organization needs to see the disclosure control. If rights liability for generated assets sits with the advertiser by contract, the AI declaration is not purely an ad-ops setting. It is a record of what was produced and how, which is the kind of documentation legal teams ask for after a complaint rather than before one.
What Happens If You Don’t Disclose
Google hasn’t announced penalties for failing to disclose AI use in ads created with third-party tools. The EU AI Act does have enforcement mechanisms, though. Article 99 allows member states to impose fines of up to €35 million or 7% of global annual turnover for prohibited AI practices. Non-compliance with transparency obligations falls under a lower tier, but fines can still reach €15 million or 3% of turnover [3].
Google’s terms of service already require advertisers to comply with applicable law. If you’re running AI-generated ads in the EU without proper disclosure, you’re risking regulatory fines and potentially violating Google’s advertising policies. That could lead to account suspension.
The Bottom Line for European Advertisers
Google’s “How this ad was made” disclosure is a compliance tool. It exists to help you meet legal obligations in markets that now require AI transparency in advertising. For advertisers in the EU, India, and New York, it’s mandatory. For everyone else, it’s still visible in My Ad Center.
The August 2, 2026 EU AI Act deadline is the date that matters. If you’re running AI-generated ads in European markets, you need to have your disclosure process in place before then. Audit your creative pipeline, document AI tool use, update your ad review workflow.
The tools make it easy to generate creative at scale. The regulations make it mandatory to be transparent about what you did. Google built the mechanism. Using it is the advertiser’s job.
Sources
- Keerat Sharma, “Expanding AI transparency in ads,” Google Ads & Commerce Blog, July 9, 2026. https://blog.google/products/ads-commerce/google-ads-ai-transparency-labels/
- Brooke Osmundson, “Google Ads Requires Disclosure For AI-Generated Content,” Search Engine Journal, July 13, 2026. https://www.searchenginejournal.com/google-ads-requires-disclosure-for-ai-generated-content/581925/
- European Union, “Regulation (EU) 2024/1689 — Artificial Intelligence Act,” Official Journal of the European Union, July 12, 2024. https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32024R1689
- “Google Ads Terms of Service Updated For AI Changes After 8 Years,” PPC News Feed, June 3, 2026. https://ppcnewsfeed.com/ppc-news/2026-06/google-ads-terms-of-service-updated/
- Matt G. Southern, “YouTube Now Auto-Detects AI Content, Labels It For Viewers,” Search Engine Journal, May 27, 2026. https://www.searchenginejournal.com/youtube-now-auto-detects-ai-content-labels-it-for-viewers/576015/
- Alejandro Borgia, “Announcing the launch of the new Ads Transparency Center,” Google Ads & Commerce Blog, March 29, 2023. https://blog.google/innovation-and-ai/technology/ads/announcing-the-launch-of-the-new-ads-transparency-center/
- “Google tests light gray background for Sponsored results,” PPC News Feed, June 10, 2026. https://ppcnewsfeed.com/ppc-news/2026-06/google-tests-light-gray-background-sponsored-results/
