What exactly did Google change about AI labeling in July 2026?

In July 2026, Google introduced a mechanism that lets advertisers add text or visual labels directly within image and video ad creatives that were generated or modified using AI, and it rolled the associated AI label setting out gradually across the month. Google presents the feature as a way to help advertisers comply with emerging AI-transparency regulations — it specifically references disclosure requirements in the European Union, India and New York that can apply when ad creative contains AI-generated or AI-edited assets. The essential point is that this is an enabling control, not a new restriction: Google is not prohibiting AI-generated creative, it is providing a sanctioned way to disclose it. That matters because, before this change, adding a disclosure overlay or watermark to a creative could itself conflict with Google's policies that prohibit text overlays and watermarks. Under the new approach, labels applied through Google's supported methods will not be treated as violating those overlay and watermark policies, so advertisers can disclose AI involvement without tripping a separate creative rule. There are two advertiser-facing routes plus an automatic one. Advertisers can add the label manually to the creative asset themselves, or they can turn on Google's AI label setting and let the platform apply the label; separately, Google Ads may automatically apply labels to assets that were created using Google's own AI tools. The capability is available for image and video assets across Google Ads, Display & Video 360, Campaign Manager 360, Merchant Center and Ads Editor, which means it is a cross-surface feature rather than a single-product one. Two limits define how advertisers should treat it. First, Google states that using the AI label setting does not guarantee compliance with any specific regulation and advises advertisers to seek their own legal guidance — the label is a mechanism, not a legal safe harbour. Second, political advertisers must continue to use the separate 'Altered or synthetic content' checkbox for election-ad disclosures rather than relying on this general setting. Taken together, the change gives advertisers a practical, policy-safe way to disclose AI-generated creative while leaving the responsibility for meeting the underlying laws with the advertiser. For the platform baseline see the Google Ads policy guide , and pre-check creative with the AI Compliance Audit . The organizing principle is that Google added a sanctioned, policy-exempt way to label AI-generated image and video creative, not a guarantee of legal compliance.

How do advertisers actually apply the AI label to a creative?

Advertisers have two direct routes to apply an AI label, plus an automatic path for assets built with Google's own tools, and choosing the right one depends on how a team produces and manages its creative. The first route is manual: the advertiser adds the label — text or a visual mark — to the image or video asset itself before it is uploaded. This suits teams that produce creative externally and want full control over exactly where and how the label appears. The second route is the AI label setting: the advertiser turns on Google's setting and the platform applies the label, a capability Google rolled out gradually across July 2026. This suits teams that prefer a consistent, platform-managed label rather than adding one by hand to every asset. The third path is automatic — Google Ads may itself apply labels to assets that were created using Google's own AI tools, so creative generated inside Google's ecosystem can be labeled without a separate step. The feature is available for image and video assets across Google Ads, Display & Video 360, Campaign Manager 360, Merchant Center and Ads Editor, so the same labeling discipline should follow an asset wherever it runs. A key practical benefit is that labels applied through these supported methods will not be treated as violating Google's policies that prohibit text overlays and watermarks; that exemption is what makes disclosure workable, because advertisers no longer have to choose between transparency and a creative that passes overlay checks. The right workflow is to identify which assets were generated or modified using AI, decide on a labeling method that fits your production process, and apply it consistently across every surface where the asset appears — an asset reused across Google Ads, DV360 and Merchant Center should carry consistent disclosure everywhere. It is also worth avoiding accidental double-labeling: if Google applies an automatic label to a Google-AI asset, adding a manual one on top may be redundant. Before scaling AI-generated creative, run it through the AI Compliance Audit to catch disclosure and policy gaps, and check copy alongside creative with the keyword risk checker . The organizing principle is that advertisers can label manually, via Google's setting, or automatically for Google-AI assets, and should apply the chosen method consistently across every surface an asset runs on.

Which regulations is this AI-labeling feature meant to address?

Google explicitly ties the AI-labeling feature to emerging AI-transparency regulations, naming disclosure requirements in the European Union, India and New York as the kinds of rules the label is meant to help advertisers address. The broader context is that lawmakers in multiple jurisdictions are moving toward mandatory disclosure of AI-generated or AI-altered content, and platforms are building tools that make it possible to comply without breaking their own creative policies. In the European Union, transparency obligations around AI-generated and manipulated content are advancing as part of a wider push to ensure audiences can identify synthetic media, reinforcing the direction toward clear labeling. In India, disclosure expectations for AI-generated or edited content form part of the same movement Google references. In New York, disclosure requirements can apply to AI-generated performers and content, an area addressed by dedicated state legislation. The common thread across all three is a single principle expressed through different legal texts: people should be able to tell when what they see was created or altered by AI. What makes this challenging for advertisers is that AI-transparency law is fragmenting by jurisdiction even as it converges on that shared principle — the precise triggers, wording and scope differ from market to market, so an approach that satisfies one jurisdiction may not automatically satisfy another. That fragmentation is exactly why a platform-level labeling mechanism is useful: it gives advertisers a consistent, sanctioned way to disclose AI involvement across markets even though the underlying laws are not uniform. It is important, though, not to over-read the feature: Google's label is a way to disclose, not a determination that any particular law is satisfied, and advertisers running cross-border campaigns must still map which rules apply where. For the EU framework specifically, see the EU DSA compliance guide ; for the state-level disclosure trend, see the California AI Transparency Act guide and the New York synthetic-performer disclosure guide . Track how these obligations evolve on the Policy Change Tracker . The organizing principle is that the feature addresses a fragmenting set of AI-transparency laws — in the EU, India and New York among others — that converge on requiring disclosure of AI-generated or altered content.

Does turning on the AI label setting make my campaign compliant?

No — turning on Google's AI label setting does not make a campaign compliant with any specific regulation, and this is the single most important limit to understand about the feature. Google states directly that using the AI label setting does not guarantee compliance with any specific regulation, and it advises advertisers to seek their own legal guidance. The label is a mechanism for disclosure, not a legal safe harbour, and treating it as one is a mistake that could leave an advertiser exposed. The reason is straightforward: AI-transparency laws differ by jurisdiction in what they require, when they apply, and how disclosure must be made. A platform setting provides a consistent way to apply a label, but it cannot itself determine whether that label, in that form, in that market, satisfies the particular legal obligation that applies to a given campaign. Compliance depends on facts the platform does not control — which laws apply to the advertiser and audience, how the AI was used, what the specific regulation demands, and whether additional steps beyond a label are required. So the correct posture is to use the label as one component of a broader disclosure process rather than as the whole of it. That process should include mapping which AI-transparency laws apply to each target market, documenting how AI was used in each creative, applying the label consistently across every surface the asset runs on, and seeking qualified legal guidance where the exposure is material. There is also a specific carve-out to respect: political advertisers must continue to use Google's separate 'Altered or synthetic content' checkbox for synthetic or digitally altered content in election ads, rather than relying on the general AI label setting — so for election advertising the general feature is not the right tool at all. The measured reading is that the label genuinely helps advertisers disclose AI-generated creative in a policy-safe way, but the responsibility for meeting the actual laws remains squarely with the advertiser. To reduce the chance of a disclosure or policy gap slipping through, pre-check creative and campaigns with the AI Compliance Audit , review the Google Ads policy guide , and confirm current requirements against official Google sources. The organizing principle is that the AI label setting is a disclosure mechanism that does not guarantee regulatory compliance, so advertisers must still meet the underlying laws themselves.

How does AI labeling interact with Google's rules on overlays and watermarks?

AI labeling interacts with Google's overlay and watermark rules through a deliberate exemption: labels applied through Google's supported AI-labeling methods will not be treated as violating the policies that otherwise prohibit text overlays and watermarks on creative. This exemption is what makes the whole feature practical, and understanding it resolves a tension that would otherwise make disclosure difficult. Google's advertising policies have long restricted text overlays and watermarks on certain creative to keep ads clean and consistent and to prevent misleading or cluttered presentation. Absent a carve-out, an advertiser trying to do the responsible thing — adding a visible disclosure that a creative was AI-generated or AI-modified — could find that the very act of adding that disclosure conflicted with the overlay and watermark rules. That would put transparency and policy compliance at odds, discouraging exactly the disclosure that AI-transparency laws are pushing toward. The July 2026 change removes that conflict. Because labels added through the manual method, the AI label setting, or Google's automatic labeling of Google-AI assets are exempt from the overlay and watermark policies, advertisers can disclose AI involvement without worrying that the disclosure itself will fail a creative check. In practice, this means advertisers should apply the AI label using one of the supported methods rather than improvising their own overlay, since a self-made overlay that is not part of Google's supported labeling might not receive the same treatment. It also means the label should be applied consistently across the surfaces where the feature is available — Google Ads, Display & Video 360, Campaign Manager 360, Merchant Center and Ads Editor — so that the same asset carries a compliant, policy-exempt disclosure everywhere it runs. Advertisers should still avoid redundant or conflicting marks; if Google automatically labels a Google-AI asset, adding another overlay manually may be unnecessary. The broader lesson is that platforms are adapting their creative rules to accommodate AI disclosure, and advertisers benefit from using the sanctioned path rather than a workaround. To keep creative and copy aligned with both the labeling feature and the surrounding policies, use the AI Compliance Audit and review the Google Ads policy guide . The organizing principle is that Google exempts its supported AI labels from overlay and watermark policies, so advertisers should disclose AI creative through the sanctioned methods rather than an improvised overlay.

What should advertisers do about AI labeling for election and political ads?

For election and political ads, advertisers should not rely on the general AI label setting and must instead continue to use Google's dedicated 'Altered or synthetic content' checkbox for synthetic or digitally altered content — this is an explicit carve-out in the July 2026 change, and getting it wrong in the sensitive context of election advertising carries heightened risk. The general AI-labeling feature was introduced to help advertisers disclose AI-generated image and video creative across ordinary campaigns, but election advertising sits under a stricter, separate disclosure regime that Google has maintained specifically for altered or synthetic political content. The practical instruction is therefore twofold. First, keep election ads on the separate mechanism: when a political or election ad contains synthetic or digitally altered content, the advertiser must use the 'Altered or synthetic content' checkbox rather than treating the general AI label setting as sufficient. Second, do not assume the two systems are interchangeable — the existence of a general labeling feature does not relax or replace the election-ad requirement, and an advertiser who applies only the general label to a political ad may fail to meet the applicable election-ad disclosure obligation. This separation reflects a broader reality that election advertising is among the most closely scrutinised categories across platforms and jurisdictions, with dedicated rules around transparency, authorisation and synthetic content. Advertisers in this space should treat disclosure as a compliance-critical step, confirm the current requirements for altered or synthetic election content against official Google sources, and recognise that political-ad rules frequently change and vary by country. More generally, the election carve-out is a reminder of the feature's overall limit: the general AI label is a disclosure mechanism, not a guarantee of compliance, and specialised categories like political advertising have their own obligations that override the general approach. Advertisers running both ordinary and political campaigns should build a process that routes each ad to the correct disclosure mechanism, documents how synthetic content was used, and seeks legal guidance where the exposure warrants it. Track political-advertising and synthetic-content developments on the Policy Change Tracker , and pre-check campaigns with the AI Compliance Audit . The organizing principle is that election and political ads must continue to use Google's separate 'Altered or synthetic content' checkbox, not the general AI label setting, for synthetic or digitally altered content.

Google Ads AI Labeling 2026: Image & Video Disclosure

Quick Answer

In July 2026, Google introduced a way for advertisers to add text or visual labels directly within image and video ad creatives that were generated or modified using AI, rolling the AI label setting out gradually across the month. Google frames the feature as a way to help advertisers comply with emerging AI-transparency regulations — it specifically references disclosure requirements in the European Union, India and New York that can apply when ads contain AI-generated or edited content. Advertisers have two paths: add the labels manually to their creative assets, or use Google's AI label setting; either way, labels applied through these methods will not be treated as violating Google's policies that otherwise prohibit text overlays and watermarks, and Google Ads may also apply labels automatically to assets created with Google's own AI tools. The feature applies to image and video assets across Google Ads, Display & Video 360, Campaign Manager 360, Merchant Center and Ads Editor. Two limits matter: Google states plainly that using the AI label setting does not guarantee compliance with any specific regulation, so advertisers should seek their own legal guidance; and political advertisers must continue to use the separate "Altered or synthetic content" checkbox for election-ad disclosures rather than relying on this setting. The practical takeaway is that this is an enabling control, not a blanket compliance solution — it gives advertisers a sanctioned mechanism to disclose AI-generated creative, but the responsibility for meeting the underlying laws remains with the advertiser. Review the platform baseline in the Google Ads policy guide, track changes on the Policy Change Tracker, and pre-check creative against platform and legal standards with the AI Compliance Audit.

Google Ads AI Labeling Requirements in 2026: Labeling AI-Generated Image and Video Creatives

What Google Announced

In July 2026, Google introduced a mechanism that lets advertisers add text or visual labels directly within image and video ad creatives that were generated or modified using AI. The capability rolled out gradually across the month, and Google presents it as a way to help advertisers comply with the growing set of AI-transparency rules that require disclosure when advertising contains AI-generated or edited assets.

The change is best understood as an enabling control rather than a new prohibition. Google is not banning AI-generated creative; it is giving advertisers a sanctioned way to disclose it. That distinction matters, because until now adding an overlay or watermark to a creative could itself run into Google's policies against text overlays and watermarks. The new labels are exempted from those policies, so advertisers can disclose AI involvement without tripping a separate creative rule.

"Advertisers can add labels to image and video creatives that were generated or modified using AI to help comply with emerging AI transparency regulations.
— Google Advertising Policies Help, AI labeling update (July 2026)"

For advertisers, the significance is practical rather than dramatic. The change does not expand or restrict what creative may run; it removes a friction point that previously made responsible disclosure awkward, and it signals that Google expects AI-generated creative to be identifiable where transparency rules apply. Treating the label as a routine part of creative production — decided at the point an asset is made rather than retrofitted later — is the way to get consistent, low-effort disclosure from it.

This guide explains the two ways to apply the label, where it is available, the regulations that prompted it, and — importantly — the limits Google itself places on what the label achieves. For the broader platform baseline see the Google Ads policy guide, and for the wider disclosure trend see the California AI Transparency Act guide.

The Two Ways to Add an AI Label

Google gives advertisers two routes to disclose AI-generated or AI-modified creative, plus an automatic path for assets built with Google's own tools. Understanding which route fits your workflow avoids both under-labeling and accidental double-labeling.

Manual Versus Setting-Based Labels

Method	How it works	Best for
Manual label	The advertiser adds the label text or visual mark to the creative asset itself before upload	Teams that produce creative externally and want full control of placement
AI label setting	The advertiser turns on Google's AI label setting so the platform applies the label, rolled out gradually through July	Teams that prefer a platform-managed, consistent label across assets
Automatic (Google AI tools)	Google Ads may automatically apply labels to assets created using Google's own AI tools	Advertisers generating creative inside Google's tools

Crucially, labels added through any of these methods will not be treated as violating Google's policies that otherwise prohibit text overlays and watermarks on creative. That exemption is what makes disclosure practical: advertisers no longer have to choose between transparency and a clean creative that passes overlay checks. Before you scale AI-generated creative, run it through the AI Compliance Audit to flag disclosure and policy gaps early.

Where the Labels Apply

The labeling capability is not limited to a single product. Google has made it available across its advertising and commerce surfaces for image and video assets, which means advertisers running multi-product campaigns can apply a consistent disclosure approach.

Products and Formats in Scope

Surface	Asset types	Note
Google Ads	Image and video	Core surface for the manual and setting-based label
Display & Video 360	Image and video	Programmatic buyers can apply labels to eligible assets
Campaign Manager 360	Image and video	Included in the same labeling rollout
Merchant Center	Image and video	Commerce assets covered
Ads Editor	Image and video	Bulk workflow support

Because the capability spans these products, advertisers should treat AI labeling as a cross-surface practice rather than a one-product task: an asset generated once and reused across Google Ads, DV360 and Merchant Center should carry consistent disclosure everywhere it runs. Operationally, that breadth is a reason to assign clear ownership — when one team generates an asset and others deploy it across different surfaces, someone needs to confirm the label travels with it, because a label applied in one place does not automatically imply disclosure elsewhere if the asset is re-uploaded. The safest default is to label at the source and verify on each surface. For teams coordinating creative across platforms, the same discipline applies on other networks — see how disclosure obligations are converging in the New York synthetic-performer disclosure guide.

The Regulations Behind the Change

Google explicitly ties this feature to emerging AI-transparency regulations, naming disclosure requirements in the European Union, India and New York. The label exists because lawmakers in multiple jurisdictions are moving toward mandatory disclosure of AI-generated or AI-altered content, and platforms are building tools that make it possible to comply without breaking their own creative rules.

The Disclosure Landscape

European Union: transparency obligations around AI-generated and manipulated content are advancing, reinforcing the direction toward clear labeling of synthetic media.
India: disclosure expectations for AI-generated or edited content are part of the same broader movement Google references.
New York: disclosure requirements can apply to AI-generated performers and content, a theme covered in dedicated state legislation.

The common thread is that AI-transparency law is fragmenting by jurisdiction while converging on a single principle: audiences should be able to tell when what they see was made or altered by AI. Advertisers running cross-border campaigns cannot assume one market's approach applies everywhere, which is why a platform-level labeling mechanism is useful — it provides a consistent way to disclose even as the underlying laws differ. Track how these rules evolve on the Policy Change Tracker, and for the EU framework see the EU DSA compliance guide.

What the Labels Do Not Do

The most important part of Google's announcement is the limit it places on the feature. Google states directly that using the AI label setting does not guarantee compliance with any specific regulation, and it advises advertisers to seek their own legal guidance. This is not boilerplate — it defines how advertisers should treat the tool.

Two Boundaries to Respect

No compliance guarantee: the label is a mechanism, not a legal safe harbour. Turning it on does not mean a campaign satisfies EU, Indian, New York or any other disclosure law; advertisers remain responsible for meeting the actual requirements that apply to them.
Election ads are separate: political advertisers must continue to use Google's dedicated "Altered or synthetic content" checkbox for synthetic or digitally altered content in election ads, rather than relying on this general labeling setting.

The practical reading is that the label helps advertisers disclose, but the obligation to comply sits with the advertiser. A team that assumes the setting resolves its legal exposure has misread the tool. The right posture is to use the label as one part of a disclosure process that also includes understanding which laws apply to each market, documenting how AI was used in creative, and confirming requirements with qualified counsel where the stakes warrant it. To pressure-test copy and claims alongside creative disclosure, use the keyword risk checker, and confirm the current text of the policy against official Google sources.

Advertiser AI-Labeling Checklist

[ ] Identified which image and video creatives were generated or modified using AI
[ ] Chosen a labeling method — manual, AI label setting, or automatic for Google-AI assets
[ ] Confirmed labels are applied consistently across Google Ads, DV360, Campaign Manager 360, Merchant Center and Ads Editor
[ ] Verified that labels added this way are exempt from overlay and watermark policies
[ ] Mapped which AI-transparency laws (EU, India, New York, others) apply to each target market
[ ] Understood that the setting does not guarantee compliance with any specific regulation
[ ] Kept election ads on the separate "Altered or synthetic content" checkbox
[ ] Documented how AI was used in each labeled asset for internal records
[ ] Sought legal guidance where AI-transparency exposure is material
[ ] Confirmed the policy's current wording against official Google sources

Don't miss the next policy change.

Create a free account — track every policy change across 8 platforms, get instant alerts, and access every free compliance tool. Or try our Keyword Risk Checker first.

Create Free Account

Report Keywords — Run AI Compliance Audit

#Google Ads#AI Labeling#Synthetic Media#Disclosure Rules#Ad Compliance#Display & Video 360#AI Transparency#Creative Compliance#Advertisers#2026 Policy#Compliance Guide 2026#Brand Safety

Share This Report

Tweet Share

YouTube Advertiser-Friendly Guidelines in 2026: Content Categories, Self-Certification, Inventory Modes and Brand Suitability

YouTube's advertiser-friendly content guidelines decide which videos earn ads, while inventory modes and content exclusions decide where advertisers' ads run — two linked systems that define brand safety in 2026.

May 2026 Enforcement Digest: Who Got Banned, Fined or Paused Across 8 Platforms

May 2026 was the heaviest enforcement month of the year so far. Every advertiser-facing action across the eight major platforms — bans, fines, pauses, and the patterns brands need to read before June.

YouTube Auto-Dubbing for Ads May 2026: Multi-Language Reach, Voice-Clone Disclosure & Ad Council Concerns

YouTube extended AI auto-dubbing to Video Action Campaigns in May 2026 — voice-clone mechanics, disclosure obligations, talent union pushback, and the regulated-vertical translation risks.

Platforms

Industries

Knowledge

Blog

Google Ads AI Labeling Requirements in 2026: Labeling AI-Generated Image and Video Creatives