
On November 11, 2026, Google clarified that ads may not use AI to impersonate public figures, even when the ad discloses the AI. A public figure here means a famous or well-known person, such as a government official, entertainer, athlete, online influencer or prominent businessperson.
If you make ad images, voice-overs or video with AI, that rule applies to your ads. A voice that sounds like a local TV host, a face that looks like a popular YouTuber or an "athlete" praising your product can all put an ad at risk. Google's policy page says suspension "occurs upon detection and without prior warning."
Google gave no start date and says the clarification leaves enforcement of the policy as it was, so treat the rule as already in force. AI-made ads are still allowed, and Google still lets advertisers label them. What matters first is whether any ad of yours makes a real, well-known person appear to endorse, use or sponsor your business.
Key Takeaways
Google's policy says ads that impersonate public figures using AI are not allowed "even if they disclose that they were generated or edited using AI."
The policy page says suspension occurs upon detection and without prior warning, so fix a risky face or voice before it runs, not after a notice.
The rule targets impersonation of real, well-known people. Google's July 9, 2026 page lets advertisers add AI labels to image and video creatives.
Find out in ten minutes whether any of your ads are at risk
Google's policy page on unacceptable business practices lists its own examples. They include a singer's name or image used without permission to suggest happy clients of an investment service, a professional athlete falsely claimed to sponsor a weight loss product, and AI-generated videos of popular YouTubers suggesting they endorse a "Get rich quick" scheme. Google's change log entry for November 11 carries the rule itself, without the examples.
The pattern in all three is the same: a real, recognizable person appears to back a product they never backed. That is the check to run on your own creative. A look-alike, a sound-alike or an AI avatar built from a famous person's features can fall into it, whether or not your ad says the person's name.

Keep making AI ads without putting your account at risk
The new wording is not a ban on AI-made ads. Google's AI labels page, dated July 9, 2026, says advertisers can add "text or visual labels directly within image and video ad creatives." It adds that Google Ads may also label certain assets made with Google's own AI tools automatically. Google also says the label setting is no guarantee of compliance with specific regulations.
So, the labels and the impersonation rule do different jobs. A label tells viewers an asset was made with AI. Only the absence of an impersonated person keeps the ad inside the rule. Background art, resized photos, a cartoon mascot and an AI voice that is plainly nobody in particular are a different question from a voice that sounds like a famous person selling your product. If you want to see how AI-written ad text looks before it runs, our post on previewing AI-written ad text in Google Ads covers that side of it.
Where a character is close to a real person, ask whether a viewer could read the ad as that person backing you. If the answer is yes or maybe, replace the asset, or ask Google Ads support about that specific creative before you rely on it.
Clean up your creative in the order that protects you fastest
Work through this in the order an owner actually does it, starting with a list of what is live and what is saved.
- List every ad and asset that shows a recognizable face, voice or name. Include AI avatars, voice clones and testimonial-style videos.
- For each one, check whether the person appears to endorse, use, sponsor or be connected with your business. A disclosure changes nothing about that check.
- Note which tool made or edited the image, voice or video.
- Keep any written consent, contract or public statement that confirms a real relationship.
- Replace anything you cannot support, then look for the same asset in other campaigns and asset libraries.
Ten-minute creative check
Any AI-made or AI-edited face that resembles a known person.
Any voice-over or voice clone that sounds like a known person.
Any ad copy that uses a well-known person's name or quotes them.
Any video or quote that shows a famous or well-known person backing you.
Written consent, a contract or a public statement for each real relationship.
Google's policy page says to "make sure you have consent from any public figures." Consent and the AI sentence are separate questions. The page asks for proof of the relationship in an appeal, so before you rely on consent for an AI likeness, ask Google Ads support about that specific creative. If you do have a real relationship, keep the paperwork where you can find it, and use a real recording of the person instead of an AI copy.

Replace risky creative with something you own
The safest replacements are original talent, licensed talent, fictional characters and creative with no identifiable person in it. For a small business, original talent often means you, your staff or a customer who agrees to appear. A short video of the owner talking to the camera costs little and cannot be mistaken for a celebrity.
If the risky asset is a long video, a voice-over or a testimonial spot, our video marketing work covers replacing it with original footage and voice. For one unsupported AI image or voice, deleting it or swapping in an original of your own may be the entire job.

Know what happens if an ad or account breaks the rule
Google's policy page puts public-figure impersonation among egregious violations. It says suspension "occurs upon detection and without prior warning," that the advertiser is not allowed to advertise with Google Ads again, and that it reinstates such accounts only in exceptional cases. Google's suspension page says it will suspend accounts for egregious violations "immediately without prior warning."
Plan as if no warning will come. If you want the longer story of what a suspension looks like from the inside, our post on a suspended Google Ads account walks through it.
| Situation | What Google says |
|---|---|
| AI impersonation of a public figure | Egregious violation; suspension "occurs upon detection and without prior warning" |
| Repeat violations of other policies | Warning at least 7 days before suspension; first violation no strike; first and second strikes bring holds of 3 and 7 days; third strike leads to suspension |
| Appeal of an ad decision | Each ad is limited to 3 appeals; wait at least 24 hours between appeals on the same ads or campaigns |
| Ad decision older than 6 months | From July 21, 2026, in-account appeals are not available; contact Google Ads support |
| Suspended account | Use the Contact Us link in the notification; at least 6 months from the suspension to appeal |
Check your audiences too, since Google changed that rule the same day
Google posted a second change on November 11, 2026. Its Personalized Advertising Policy update says advertisers are not allowed to target users based on an interest in categories prohibited in Google's ad network.
Google's targeting page links to the list of those categories. It says a flagged ad can be fixed by editing the targeting, editing or removing the content, or appealing, and that an edited ad is automatically reviewed again. It adds that removing ads can help prevent suspension for repeated violations. Open your audience segments, and edit any that aim at an interest in a prohibited category.
Appeal once the problem is fixed, and keep the proof close
Fix the asset first, then appeal. In Google Ads, go to Tools, then Troubleshooting, then Policy manager. Open the Policy issues tab, choose Appeal, and pick the reason: "Dispute decision" or "Made changes to comply with policy." You can also select the disapproved ads, choose Edit, then "Appeal policy decision." Google's appeals page says each ad is limited to 3 appeals, and that after 3 unsuccessful appeals you contact support.
For a real relationship with a public figure, the policy page asks for details on that relationship and proof such as a public statement, and says proof can include contracts and written agreements. A minor edit to ad text or the final URL triggers a new crawl, which is worth knowing if your fix changed only the wording.
If the whole account is suspended, use the "Contact Us" link in the account notification. The suspension page says some advertisers must complete advertiser verification first, which is covered in our post on what Google's advertiser verification asks for. No one can promise an appeal succeeds, so build your case on what you changed.
How these rules reached this point
- 1
July 9, 2026
Google's AI labels page lets advertisers add text or visual labels to image and video creatives.
- 2
July 21, 2026
In-account appeals stop being available for decisions older than 6 months.
- 3
November 11, 2026
Google's change log clarifies that ads may not use AI to impersonate public figures, even with a disclosure.
- 4
November 11, 2026
Google clarifies the Personalized Advertising Policy on targeting by prohibited-category interests.
Decide whether to do this yourself or bring in help
You can run the first pass alone. The ten-minute check above needs only your Google Ads login and your creative folder. If the only problem is one unsupported AI voice or image, removing or replacing it may be the whole job, and no outside help is needed.
Help is worth paying for in a few situations: many campaigns, assets duplicated across them, several markets, an agency handoff, a prior suspension, unclear licensing, or a campaign that cannot afford downtime. In those cases, an audit that traces every copy of an asset saves the guesswork. No one can promise approval or reinstatement, and anyone who does is selling something Google controls.

Can your AI ad run under the new rule?
Pick an answer to begin.
1. An ad uses an AI voice that sounds like a well-known athlete praising your product, and the ad says "AI-generated." What does Google's policy say?
2. Can you still put AI labels on image and video ads?
3. An ad is disapproved and you have fixed the asset. How many times can you appeal that one ad?
Frequently Asked Questions About the Google Ads AI Impersonation Policy
Can I use an AI voice that sounds like a celebrity in a Google ad?
Google says ads that impersonate public figures using AI are not allowed, even if the ad discloses the AI. A sound-alike made to suggest a famous person backs your business is the situation the rule covers.
Does an AI label make a look-alike ad acceptable?
No. Google's wording says AI impersonation of a public figure is not allowed, disclosure or not. Labels remain an option for AI-made creative that impersonates no one.
Who counts as a public figure under Google's policy?
Google describes public figures as famous or well-known people, including government officials, entertainers, athletes, online influencers and prominent businesspeople.
Will Google suspend my account or just reject the ad?
The policy page says suspension occurs upon detection and without prior warning, and it treats violations as egregious. A single ad can also be disapproved, so fix any risky asset before it runs.
Can I get permission from the person and keep using the AI version?
The policy page asks you to have consent from any public figures and to supply proof of the relationship in an appeal. Consent and the AI rule are separate questions, so ask Google Ads support about the specific creative before you rely on it, or use the person's own recording.
Can I still target audiences by interest?
Yes, with one limit. Google says advertisers are not allowed to target users based on an interest in categories prohibited in Google's ad network, so check your segments against that list.
What This Means for You
Google says ads that impersonate public figures using AI are not allowed, and that a disclosure changes nothing. Suspension can come without prior warning, and AI-made ads with no real person's likeness in them are still allowed, with labels available. Check every face, voice and name, then your audiences, then keep proof for any real relationship.
Doing the check now means your next campaign launches without a surprise notice, and your account stays open for the ads that bring in work. A short list of who is in each ad and who agreed to it is a habit that keeps paying off.
If your account is large or you want a second set of eyes, Web Leveling can audit your ads and assets, review Policy manager, correct what needs fixing and support an appeal, through our Google Ads management work. We tell you when the answer is simply to delete one asset. We work with small and medium businesses across the country and overseas. Contact us about a Google Ads policy check and we will look at what is running.
Terms
Terms used in this post
Tap a term to see what it means.
Public figure. In Google's policy, a famous or well-known person, including government officials, entertainers, athletes, online influencers and prominent businesspeople.
Impersonation. Using AI to make an ad's face, voice or video appear to be a real person who has not agreed to appear in it.
Egregious violation. A policy area where Google can suspend an account immediately, without a prior warning.
Policy manager. The Google Ads page, under Tools and Troubleshooting, that lists policy issues and lets you appeal.
AI label. Text or a visual mark inside an image or video ad showing it was made with AI.
Prohibited category. A category of content that Google's ad network does not allow, and that you cannot target users by interest in.




