Checked September 3, 2026
Google AI ad labels: what advertisers must check
Record how each image or video was made, which label rule applies, who approved it, and what the buyer can see in the live ad.
On this page
Direct answer
Google now has more than one AI disclosure path.
Google says ads created with its own generative tools can receive an automatic disclosure. It also provides an AI content label setting for eligible image and video assets made or edited with outside tools. Google-generated media can carry machine-readable signals such as SynthID or C2PA metadata.
Those mechanisms are not the same. An internal label setting, machine-readable provenance, and a buyer-facing disclosure may appear in different places. Google also says using its setting does not guarantee compliance with every law. The advertiser still owns the truth of the creative and the approval record.
Use the advertising service when the issue needs account work, the agency guide to check ownership and records, and Learn for other current marketing guidance.
Four paths
Identify which disclosure mechanism applies
| Mechanism | What it means | Advertiser check |
|---|---|---|
| Google-created asset | Google says ads created with its generative advertising tools receive an automatic disclosure. | Keep the prompt or brief, source inputs, selected output, approval, campaign, and live-ad evidence. |
| Advertiser-selected label | An eligible image or video created or edited with an outside generative tool can be marked through Google’s label control. | Check the asset status, campaign use, region, safe area, and what the buyer sees. |
| Machine-readable provenance | Google documents SynthID and C2PA signals for some generated media. | Do not treat metadata alone as the same thing as a visible disclosure in every placement. |
| Separate policy or legal duty | Political content and local law can create other disclosure requirements. | Review the applicable rule and creative. Google’s product setting is not a universal legal safe harbor. |
Asset record
Keep nine facts with every generated asset
- Original source file and its owner or license.
- Tool and account used to create or edit the asset.
- Prompt, brief, reference files, and material edit notes.
- What is real, generated, removed, moved, or changed.
- Product, person, location, result, and testimonial accuracy check.
- Human approver and approval date.
- Google asset ID, campaign, and label status.
- Regions and placements where the asset can serve.
- Screenshot or recording of the buyer-facing ad after launch.
Approval decision
Ask what the image makes a buyer believe
Person
Does a generated worker, customer, expert, or spokesperson look like a real person connected to the company?
Place
Does a generated office, storefront, job site, or service area imply a physical location that does not exist?
Product
Does the image change size, finish, contents, condition, performance, stock, or included features?
Result
Does the creative imply a before-and-after result, testimonial, savings, outcome, or speed the business cannot support?
A polished image can still make a false claim. Label review does not replace the ordinary advertising-policy and factual review.
Live check
The account setting is not the final receipt
01
Check the library
Record the AI label status for the live asset.
02
Check the campaign
Confirm which asset version is eligible to serve.
03
Check the ad
View the result in the buyer-facing surface and region.
04
Keep the proof
Save the date, placement, screenshot, and approval record.
Common questions
Answers without a sales pitch
Does Google automatically label every AI-made ad?
Google documents automatic disclosures for ads created with its own generative advertising tools. Outside-tool assets can require an advertiser-selected label. The exact result depends on asset type, region, placement, and current product rules.
Is C2PA metadata the same as a visible ad label?
No. C2PA is machine-readable provenance. A buyer-facing disclosure is a presentation layer. One does not prove the other appears in every ad surface.
Does the Google setting guarantee legal compliance?
No. Google says its setting does not guarantee compliance with specific regulations. The advertiser must review the law, policy, creative claim, and placement.
What should an agency hand back?
The business should receive source files, prompts or briefs, edit notes, approvals, asset IDs, label decisions, campaign mappings, and live-ad proof.
Primary sources
What the platform documentation supports
These sources support the facts on this page. They do not disclose complete ranking systems or guarantee a recommendation, citation, click, lead, or appointment.
- Google Ads: Expanding AI transparency in adsGoogle’s announcement and buyer-facing disclosure direction.
- Google Ads Help: AI content label settingsAsset status, setup, regional overlay, and safe-area details.
- Google Ads policy update: AI labelsGoogle’s current AI-label policy update.
- Google Ads Help: Generated imagesReview responsibility, SynthID, C2PA, and restrictions.
- Google Ads policy: Image asset requirementsImage format, text, safe-area, and label notes.
- Google Ads policy: Political contentSeparate election-ad disclosure requirements.
- Google Ads policy: MisrepresentationThe ordinary truthfulness boundary for ads and destinations.
- Google Ads policy: Editorial requirementsCreative quality and presentation rules.
- Google DeepMind: SynthIDGoogle’s description of invisible watermarking for AI content.
- C2PA: SpecificationsThe technical provenance standard Google references.
Keep creative, approval, and live proof connected.
Stan Consulting can add this record to a paid-media workflow and fix the account process around it.
Request a quote