Guide · Running AI ads
AI ad labels: what TikTok, Meta and YouTube require for AI UGC ads
Published 2026-11-13
If your ad uses an AI actor, an AI voice or an AI-generated scene, each platform has its own rule about labeling it. TikTok can reject ads that don’t disclose it. Meta adds its own label. YouTube asks at upload. On November 13, 2026 we read the current policy pages of TikTok, Meta, YouTube and Google Ads. This guide sums up what each one requires for UGC-style ads, and what a label does not fix.
The short answer
- TikTok: ads with significant AI content must be labeled. Undisclosed AI content can get the ad rejected or restricted.
- Meta (Facebook and Instagram): Meta labels AI-made or AI-edited ads itself, using its own tools and industry signals. Political and social issue ads have stricter disclosure rules.
- YouTube: realistic AI content must be disclosed when you upload the video. Google Ads has a separate checkbox rule for election ads.
- Everywhere: a label doesn’t make an AI actor a real customer. Never present AI video as a real review.
The rules side by side
| Platform | When a label applies | How it’s added |
|---|---|---|
| TikTok ads | Fully AI-generated images, video or audio, or a person shown doing or saying something they didn’t | “This ad contains AI-generated content” toggle in Ads Manager, or your own clear disclaimer |
| TikTok Spark Ads | Same idea, under TikTok’s Community Guidelines | The AI-generated content label on the creator’s post |
| Meta ads | Ads made or significantly edited with generative AI | Meta adds an “AI info” label, including when it detects third-party AI tools |
| YouTube videos | Realistic content that’s AI-made or altered | The AI use question in YouTube Studio at upload |
| Google election ads | Synthetic content that inauthentically depicts real or realistic-looking people or events | “Altered or synthetic content” checkbox in campaign settings |
Summarized from each platform’s policy pages, checked November 13, 2026. Policies change often and enforcement varies by country. This is general information, not legal advice.
TikTok: label it or risk rejection
TikTok’s ad policy on misleading and false content (last updated April 2026) allows AI-generated or significantly altered content only if it’s labeled. It says that if TikTok finds AI content that wasn’t disclosed, the ad will be rejected or restricted.
The policy gives three examples of significant AI content:
- Images, video or audio that are completely AI-generated.
- The main subject shown doing something they didn’t actually do.
- The main subject made to say something they didn’t say, using AI voice cloning.
Lighting and color adjustments, background removal and noise reduction are listed as insignificant edits. If you’re unsure which side your edit falls on, TikTok’s advice is to label it.
How to disclose: TikTok’s policy allows the AI-generated content label, or your own clear disclaimer, caption, watermark or sticker. For standard ads, Ads Manager has a toggle, “This ad contains AI-generated content”, which adds a text label. TikTok’s disclaimer guide notes that once you submit the ad, you can’t switch the toggle off unless you change the video or image. Spark Ads aren’t eligible for the toggle. They follow TikTok’s Community Guidelines on AI content instead, so the label goes on the creator’s post.
The same policy bans using a public figure’s likeness without permission, including fake endorsements, and says TikTok removes private people’s likenesses used without permission too.
Meta: the label is mostly added for you
Meta’s transparency update for ads (first published February 2025, updated June 2026) describes how its “AI info” label works:
- Ads significantly edited with Meta’s own generative AI tools get a label, either behind the three-dot menu or next to “Sponsored”.
- If those tools add an AI-generated photorealistic person, the label appears next to “Sponsored”, where more people will see it.
- Meta says it automatically detects ads made or edited with third-party AI tools through industry-standard signals, and adds “AI info” in the “About this ad” panel.
For ordinary product ads, Meta’s post describes labels it applies itself rather than a duty for you to add one. Social issue, electoral and political ads are different: Meta requires those advertisers to disclose digitally created or altered images, video and audio. Meta also notes that the experience can vary by region because of local laws.
In practice, assume an AI actor video made with an outside tool may be labeled on Meta, and make sure the ad still makes sense when it is.
YouTube and Google Ads
Ads on YouTube are usually uploaded to a YouTube channel first. YouTube’s help page on AI disclosure asks creators to disclose realistic content that’s AI-made or altered, using the AI use setting in YouTube Studio. Its examples include making it look like someone gave advice they didn’t give, and AI-generated footage of a real place.
It doesn’t require disclosure for clearly unrealistic content, minor edits such as color correction or background blur, AI help with scripts, titles or thumbnails, captions, or cloning your own voice for a voice-over. Depending on the content, the label appears on the video player or in the description. YouTube says disclosure alone won’t limit a video’s audience, but creators who repeatedly fail to disclose can face penalties.
Google Ads has its own synthetic content rule for election ads. Election advertisers must tick “Altered or synthetic content” in campaign settings when an ad inauthentically depicts real or realistic-looking people or events. For some formats, including Shorts, in-feed and in-stream ads, Google then generates the disclosure. The policy page we read covers election ads; a product ad still has to follow Google’s other ad policies, including those on misleading content.
Does AI UGC need a label?
Usually, yes. A video with an AI actor is completely AI-generated video, which is TikTok’s first example of content that must be labeled. On YouTube, a realistic AI person talking to camera is the kind of content the disclosure question is about. On Meta, the label may be added whether or not you do anything.
A label on its own isn’t the problem. What hurts an ad is a label that contradicts the script. An AI actor saying “I’ve used this every day for a month” next to an AI label looks like exactly what it is: a claim nobody made.
A label doesn’t fix a fake review
The FTC’s rule on fake reviews and testimonials covers reviews that misrepresent that they’re by someone who doesn’t exist, and names AI-generated fake reviews as an example. The FTC can seek civil penalties against businesses that knowingly break it.
So the safe way to write an AI UGC script is as a presenter, not a customer:
- Instead of “I’ve tried every serum and this is the only one that worked for me”, try “Here’s what’s in this serum and how to use it.”
- Instead of “My dog loves these”, try “Three things dog owners ask about these treats.”
- Instead of showing results “after 30 days”, show the product, the offer or the steps.
If you need a real reaction or a real result, hire a creator who actually uses the product. Our AI UGC vs real creators guide covers where that line sits.
Checklist before you launch an AI ad
- List every AI element: actor, voice, scene, product shot, music.
- On TikTok, switch on the AI-generated content toggle for standard ads, or label the creator’s post for Spark Ads.
- On YouTube, answer the AI use question when you upload the ad video.
- On Meta, preview the ad and check that it still works with an “AI info” label.
- Read the script as if the label were already there. Remove anything that sounds like a personal experience.
- Don’t use a real person’s face or voice, famous or not, without written permission.
- Keep a note of which tool made each asset, in case a platform asks.
Common mistakes
- Using a Spark Ad to skip the toggle. Spark Ads have no Ads Manager toggle, but the creator’s post still needs TikTok’s AI label.
- Assuming the rules for ordinary posts cover ads. TikTok’s ad policy has its own AI rule, and Google’s checkbox rule sits in its ad policies, not YouTube’s.
- Stripping metadata to avoid detection. TikTok says undisclosed AI content gets rejected or restricted. Hiding it adds risk without changing the rule.
- Writing testimonials for AI actors. The label can’t make a made-up experience true.
- Mixing real and AI footage without saying so. If a real creator’s face or voice is changed with AI, that is a significant edit.
Affiliate link
The link below is an affiliate link. If you sign up and pay through it, we may earn a commission at no extra cost to you. MakeUGC makes UGC-style ads with AI actors, so the labeling rules above apply to its videos too. We have not tested it ourselves. Our MakeUGC pricing guide explains the plans, credits and the US$1 trial.
Common questions
Do I have to label AI-generated ads on TikTok?
Yes, if the ad has significant AI content, such as a fully AI-generated video or a cloned voice. TikTok’s ad policy says undisclosed AI content can lead to the ad being rejected or restricted.
Does Meta label AI ads automatically?
Often. Meta says it labels ads significantly edited with its own AI tools and detects ads made with third-party AI tools through industry signals. The “AI info” label usually sits in the “About this ad” panel, or next to “Sponsored” when Meta’s tools add a photorealistic person.
Will an AI label hurt my ad’s performance?
None of the platforms publish numbers on this. YouTube says disclosure alone won’t limit a video’s audience. The bigger risk is a script that pretends to be a customer, which a label makes obvious.
Do I need to label a UGC video made by a real creator?
Not for AI, unless AI significantly changed it, for example by cloning a voice or adding scenes that didn’t happen. You still need the usual paid-partnership disclosures; see the disclosure section of our UGC vs influencer guide.
Prefer a real person on camera? Browse our creator shortlist or read the hiring guide.