TerradsTerrads
← All posts

Do AI avatar ads need a label? What Meta, TikTok and Google's rules say

TikTok makes you label any AI-generated ad. Meta and Google label it themselves or by region. All three ban faking a real person's endorsement.

On TikTok, yes: an ad with completely AI-generated video, image or audio has to carry TikTok's AI-generated content disclaimer, or a clear label of your own, and TikTok says an undisclosed one "will be rejected or restricted." On Meta, no for an ordinary commercial ad: Meta labels detected AI content itself, and you have to disclose only in ads about social issues, elections or politics. On Google, including YouTube ads, the only mandatory disclosure in its ads policy is for election ads; for everything else there is an optional AI label setting, which also puts a visible overlay on the ad in the EU, India and New York. All three ban faking a public figure's endorsement; TikTok also requires permission to use anyone's likeness.

An AI avatar, an AI-generated person who presents the product, sits right where these rules are written. A person who doesn't exist, a real person shown saying something, a cloned voice: these are the cases each platform names.

Do you have to label an AI avatar ad on TikTok?

Yes. TikTok's page About ad disclaimers in TikTok Ads Manager lists the cases where a disclaimer is optional and the one where it isn't:

Mandatory disclaimers: AI-generated, synthetic, or manipulated media. This includes images, video, or audio that are completely AI-generated or real source material that has been significantly modified by AI.

The ads policy that page points to, Misleading and false content, says the same thing as a condition. "Significantly edited media and AIGC content is allowed if" you "Apply the AIGC label, or by adding a clear disclaimer, caption, watermark, or sticker of your own." It names what counts as significant:

  • "Content that contains images, video, or audio that are completely AI-generated"
  • "Showing the primary subject doing something they didn't actually do, like dancing."
  • "Making the primary subject say something they didn't actually say, using AI voice-cloning."

And what doesn't: lighting, brightness or color changes, "Removing or modifying backgrounds", and denoising. If you aren't sure, TikTok's advice is to "use caution and label it accordingly." TikTok states the consequence: "If we identify AI-generated content that has not been disclosed, your ad will be rejected or restricted." How long that review takes, and what a rejection looks like, is in how TikTok's ad review works.

An AI person talking to camera is "completely AI-generated" video, so an ad that opens with one falls under the rule even when the rest of it is your real product photos.

Where does TikTok's AI label go, and how do you add it?

In TikTok Ads Manager it is a toggle, not something you draw into the video. How to add disclaimers to ads says to check This ad contains AI-generated content during ad creation, and the overview page says the label is "a textual disclosure label indicating that the ad includes AI-generated content", shown "at the bottom of the in-feed video for the duration of the ad." That is the same strip of the frame TikTok's interface already covers, which is one more reason to keep your caption and logo inside TikTok's safe zone (drawn on the in-feed video spec page).

Two details from the same page are easy to miss:

  • The toggle is locked once submitted: "you cannot change the toggle or unselect it unless you change the video or image." And "Any new campaign creation, including campaign duplication, will reset the AI-generated content disclaimer toggle, even if you reuse an ad that you previously used this toggle for." Duplicate a campaign and you have to switch it on again.
  • Spark Ads are excluded from the toggle. The conditions are that the ad "Is a TikTok ad" and "Is not a Spark ad", because "As Spark ads live organically on TikTok, they must follow the TikTok Community Guidelines policy for AI-generated content disclosure." That is the creator-side label set when the post is published.

TikTok's overview page also warns: "You might not see the feature for adding disclaimers if the feature is not available to you." In that case the ads policy's other route, "a clear disclaimer, caption, watermark, or sticker of your own", still satisfies the rule.

TikTok labels the creatives its Smart+ tools generate for you. For your own upload after its automatic enhancements have edited it, TikTok doesn't say.

Does an AI voice need a label on TikTok?

The two TikTok pages answer this differently, and which one applies depends on whether the ad is a Spark Ad.

The Community Guidelines, which cover Spark Ads, require a label on "AI-generated or significantly edited content that shows realistic-looking scenes or people", list "AI-generated audio mimics the voice of a real person" as needing disclosure, and say disclosure isn't needed when:

Using generic text-to-speech (TTS) narration, when the TTS isn't a recognizable voice of a known individual

The ads policy has no such carve-out. It counts "audio that are completely AI-generated" as significant, full stop. So a generic AI voiceover on a Spark Ad needs no label under the guidelines, while the same voiceover on a non-Spark ad meets the ads policy's definition. A cloned voice of a real person needs a label on both.

Does Meta require you to label an AI-generated person in an ad?

Not for a commercial ad. Meta's mandatory self-disclosure is limited to ads about social issues, elections or politics, and that is the one Meta rule that names a synthetic person. Its page About media created or edited with AI says advertisers must disclose when such an ad contains photorealistic media made or edited with AI to:

Depict a realistic-looking person that does not exist or a realistic-looking event that did not happen, or alter footage of a real event that happened.

along with depicting "a real person as saying or doing something they did not say or do." Skip it and "your ad may be removed. Repeated failure to disclose may result in penalties on the account."

For everything else, Meta labels. Its page About AI info on ads created or edited with generative AI tools says Meta uses "industry-standard detection methods, such as C2PA" to spot third-party AI content and attaches an AI info label under the three-dot menu, in About this ad. The longer version, and why there's nothing to tick, is in Meta's AI disclosure rules and how Meta detects AI.

Two exceptions bring the label out of the menu. In "the European Region, California, New York, India and Taiwan," advertisers "will have the opportunity to self-disclose that GenAI was used and if they do so, the label will appear on the ad next to the Sponsored label." And for people made with Meta's own generative tools, Meta's newsroom post says: "When these tools result in the inclusion of an AI-generated photorealistic human, the label will appear next to the Sponsored label (not behind the three-dot menu)." That second rule is about Meta's tools; a person made elsewhere and detected through C2PA gets the label in About this ad.

Do Google and YouTube ads need an AI disclosure?

Only election ads, under Google's ads policy. Google's Political content policy says "Advertisers must disclose all election ads that contain synthetic or digitally altered content" by ticking Altered or synthetic content in campaign settings. Its first in-scope example is "An ad with synthetic content that makes it appear as if a person is saying or doing something they didn't say or do." For feeds and Shorts on mobile, and in-stream, Google generates the in-ad disclosure itself; for other formats you add a "clear and conspicuous" one. The same policy says violations "won't lead to immediate account suspension without prior warning": "A warning will be issued at least 7 days prior to any suspension of your account."

Every other ad gets a setting instead of a rule. Google's Use AI content label settings and disclosures describes a per-asset choice, "Label this asset as created or edited with AI" or "Don't label this asset". A labeled asset adds a disclosure to the "How this ad was made" section of My Ad Center, reached from the three-dot menu "on ads across Search, YouTube and Discover." In three places the label moves onto the ad:

For campaigns that target the following locations, ads with assets that you designate as AI created or edited will also include visible overlays on the ads: European Union, India, New York

Google's July 2026 policy update explains why: "AI regulations in the European Union, India, and New York require that ads with certain AI-generated or edited assets include disclosures and/or labels." It also lets you put a text or visual AI label in the creative itself without breaking the overlay and watermark rules. Both pages add that the setting "doesn't guarantee your compliance with specific regulations." (What the EU rule covers is in what the AI Act asks of advertisers.)

Google can also label for you: "in regions where Google itself is legally required to do so, or when we get signals to do so from other platforms", and "In any of these cases, labels cannot be overwritten."

YouTube adds one more step for video ads. The same Google Ads page says "If you upload a video to YouTube from within Google Ads, you will be asked whether AI was used to generate or edit the video to comply with YouTube's AI disclosure policies for organic videos." Those policies, on YouTube's Disclosing use of GenAI content page, require creators to disclose realistic content that "Makes a real person appear to say or do something they didn't do" or "Generates a realistic scene that didn't actually occur", and say "Disclosing AI content alone won't limit a video's audience or impact its eligibility to earn money." That page is written for creators and doesn't mention ads.

Can you use a real person's face or voice in an AI ad?

With their permission, and the platforms differ in how far they spell that out.

TikTok is the most explicit. Its ads policy, under "Identity misuse":

We do not allow content that uses or promotes the misuse of the likeness (audio or visual) of a public figure without their permission, particularly in misleading contexts. When we become aware of likeness of private figures used without permission, we will remove such content as well.

The Community Guidelines define likeness as "A recognizable image, video, or audio representation of a person, including their face, body, voice, and gestures," and list "Using the likeness of private figures without consent" as not allowed. So on TikTok an avatar built from an employee's or a customer's photo needs their consent, not only a celebrity's.

Google treats impersonating a public figure as a scam. Its March 2024 update to the Misrepresentation policy added "enticing users to part with money or information by impersonating or falsely implying affiliation with or endorsement by a public figure, brand, or organization," and the penalty is the harshest of the three: "we will suspend your Google Ads accounts upon detection and without prior warning, and you will not be allowed to advertise with us again." The current Misrepresentation policy also bans "Manipulating media to deceive, defraud, or mislead others."

Meta covers it under Prohibited Commercial Practices, which bans "Celebrity-bait scams, which involve falsely using the name, likeness, or imagery of a public figure, together with misleading tactics to bait people into engaging with the content in an attempt to scam or defraud." The ad standard that points to it lists "unauthorized endorsements" among the practices it prohibits. Meta's policy pages don't set a consent rule for an ordinary person's likeness the way TikTok's do.

Can an AI person give a testimonial or endorse your product?

The pages draw the line at real people. TikTok's ads policy gives "Public figures being the victim of or engaging in bullying or harassing, making a false endorsement, or being endorsed" as an example of what is not allowed, and its Community Guidelines ban AI content showing "A public figure taking political stances, supporting products, or commenting on public issues they haven't actually addressed." Google's and Meta's impersonation rules above cover the same ground for public figures.

None of the three platforms' policies names an invented, AI-generated customer giving a testimonial as its own category. What does apply is each platform's rule on claims: TikTok's "Ad content and landing pages must not promise or exaggerate results concerning a product's effect", Google's ban on "claims that entice the user with an improbable result", and Meta's "Product misrepresentation: promoting products or services that materially differ from what was advertised." An AI person saying something false about your product breaks those the same way a real one would, and on TikTok the same ad also needs the AI label.

A still image ad with a made-up person in it falls under the same rules: TikTok's mandatory disclaimer names images as well as video, Meta labels the AI content it detects in any ad, and Google's AI label is set per asset. In the US, the FTC's Rule on the Use of Consumer Reviews and Testimonials makes it a violation for a business to create a consumer testimonial that misrepresents "That the reviewer or testimonialist exists." A made-up person can speak for the brand, but presenting them as a customer or a reviewer is what that rule covers.

Meta, TikTok and Google side by side

MetaTikTokGoogle (incl. YouTube ads)
Must you label an AI person in a commercial ad?No; Meta labels detected AI content itselfYes, completely AI-generated video, image or audioNo; optional AI label setting
Where the label showsAbout this ad (three-dot menu); next to Sponsored for self-disclosure in five regions, and for a photorealistic human made with Meta's own toolsText label at the bottom of the in-feed video"How this ad was made" in My Ad Center; visible overlay in the EU, India and New York
Political / election adsMandatory disclosure, incl. "a realistic-looking person that does not exist"The same AI rule as any other adMandatory "Altered or synthetic content" checkbox
What happens if you don'tPolitical ads: "may be removed", account penalties on repeat"will be rejected or restricted"Election ads: a warning at least 7 days before any suspension
Real person's likenessCelebrity-bait scams bannedPublic figures need permission; private figures' likeness without consent removedImpersonating or implying a public figure's endorsement: suspension without warning
AI voiceCovered by the political rule as "realistic sounding audio"Ads policy: AI-generated audio counts; guidelines exempt generic TTS (Spark Ads)Election rule covers audio; example disclosure "This audio was computer generated."

What to do on each platform

  1. On TikTok, switch the AI toggle on for any ad that opens with an AI person, and switch it on again after duplicating a campaign.
  2. On Meta, there is nothing to tick for a commercial ad. Delivering in the European Region, California, New York, India or Taiwan, you can self-disclose and the label moves next to Sponsored.
  3. On Google, use the AI label setting for campaigns that target the EU, India or New York; Google says those places' laws require a disclosure. Uploading the video to YouTube from Google Ads also asks the AI question.
  4. Political and election ads are a different regime on Meta and Google, with mandatory disclosure.
  5. A real person's face or voice needs their permission. TikTok requires it for everyone; Google suspends accounts without warning for faking a public figure's endorsement.

If you make the presenter in Terrads, the person is either AI-generated from a description, so they are no one in particular, or made from a photo of someone who agreed, which you confirm with a tick before they can appear in an ad. Either way the presenter is AI-generated video, so on TikTok the ad still needs its AI label.

Sources

Where a platform shows a date, it's in the list above; Google's policy pages show none. These rules change often, and Google's AI label setting was still rolling out in July 2026, so check the page before relying on any of it.

Terrads makes video ads where a person you make once opens the ad, and your real product photos carry the rest.