Are AI UGC Ads Allowed on Meta and TikTok?

TL;DRThe short answer

Yes. You can run AI UGC ads (videos with AI avatars, AI voices or AI-generated scenes that look like creator content) on both Meta and TikTok. What you cannot do is use AI to deceive: pretend an AI actor is a real customer, invent results, use a real person's face or voice without permission, or hide that a realistic video was generated when the platform requires you to say so.

Think of the rules in two layers. The first layer is specific to AI: labels and disclosure. The second layer is the normal advertising rulebook (misleading claims, testimonials, before-and-after images, landing page consistency), which applies to every ad whether a camera or a model made it. Most AI ads that get rejected fail the second layer.

DisclosureDo I need to label AI content?

On TikTok: yes, for fully or significantly AI-generated media

TikTok's ad policy says significantly edited media and AI-generated content are allowed if you disclose them, either with TikTok's AI-generated content (AIGC) label or with “a clear disclaimer, caption, watermark, or sticker” of your own. If TikTok finds undisclosed AI-generated content, the policy says the ad will be rejected or restricted.

What counts as “significantly modified by AI”? TikTok's examples include:

• Completely AI-generated images, video or audio (an AI avatar talking about your product falls here).
• Showing the main subject doing something they didn’t do.
• Making the main subject say something they didn’t say, for example with voice cloning.

Minor AI edits do not need a label: TikTok lists lighting, brightness and color adjustments, removing or changing backgrounds, and denoising as insignificant edits. If you are unsure, TikTok's own advice is to label it.

In TikTok Ads Manager the label is a toggle called “This ad contains AI-generated content” in the ad creation step (TikTok’s disclaimer help page walks through it). Two details trip sellers up:

• Spark Ads (boosted organic posts) can't use the toggle. They follow TikTok's Community Guidelines, so the AI label has to be set on the organic post itself.
• Duplicating a campaign resets the toggle, according to TikTok's help page, even if you reuse the same ad. Re-check it every time you duplicate.

On Meta: Meta labels AI ads, and political ads must self-disclose

Meta's approach is mainly automatic labelling. According to its help pages, Meta adds an “AI info” label to ads created or significantly edited with its own generative AI features (such as background or image generation) and to ads it detects as created or edited with third-party AI tools, using industry-standard signals. The label sits in the “About this ad” menu and sometimes next to the Sponsored label. Minor changes like resizing or color correction don't get it.

A separate, stricter rule applies to ads about social issues, elections or politics: there, advertisers must disclose photorealistic images, video or realistic audio created or edited with AI in certain ways. A normal ecommerce product ad is not in that category, but if your product touches a social issue, check the policy.

Claims and testimonialsCan AI avatars make claims?

An AI avatar can say anything a human spokesperson could say truthfully. It can describe the product, demonstrate a use, explain an offer and speak to a problem. The platforms judge the claim, not the face delivering it, so the usual rules apply in full:

• No exaggerated results. TikTok's policy lists examples such as instant slimming claims and creams that make wrinkles disappear in a before-and-after. If the product can't reliably do it, the avatar can't say it.
• No absolute claims you can’t back up, such as “number 1” statements. TikTok calls these out explicitly.
• No fake customer stories. An AI actor saying “I’ve used this for a year and it changed my skin” presents an experience nobody had. That is a fabricated testimonial, whatever the platform says about AI.
• No fake experts. Don't dress an avatar as a doctor, dentist or trainer to imply professional endorsement you don't have.

Outside the platforms, consumer protection law points the same way. In the US, for example, the FTC's rule on consumer reviews and testimonials bans fake reviews and testimonials, including ones attributed to people who don't exist, and AI-generated ones are covered. Other regions have similar rules on misleading advertising.

A safe way to write AI UGC scripts is to cast the avatar as a presenter, not a customer: “Here’s why this bottle doesn’t leak in a gym bag” instead of “I’ve tested every bottle and this is the only one that works.” You keep the native, creator-style feel without inventing a life story. For more on the format, see our explainer on what a UGC video ad is.

Common rejection reasonsWhat gets AI ads rejected?

Based on the policies above, these are the patterns most likely to get an AI UGC ad rejected or restricted:

Notice that only the first row is truly about AI. The rest are ordinary advertising mistakes. AI just makes them faster to produce at scale, which is why a quick review step matters more when you generate many variations. If your ads get through review but still don't sell, our guide to creative fixes for ads that don’t convert covers the next problem.

A safe processHow do I stay safe?

A repeatable process beats memorising policies. This is the workflow we recommend for any AI UGC ad:

MistakesCommon mistakes to avoid

• Assuming “everyone does it” means it’s allowed. Plenty of rule-breaking ads run for a while before they're caught. Your account carries the risk, not theirs.
• Labelling once and forgetting. Duplicated TikTok campaigns reset the AI toggle. Build the check into your launch routine.
• Letting AI write claims unchecked. AI copy tends to be confident. “Clinically proven”, “best on the market” and “results in days” often appear unprompted. Delete anything you can't prove.
• Using AI before-and-afters. Even if the product works, a generated transformation shows a result nobody achieved.
• Cloning a creator’s voice or style too closely. Recreating a competitor's popular ad beat for beat with a look-alike presenter invites likeness and IP problems.
• Ignoring category rules. Supplements, weight loss, skincare with medical claims, finance and adult products have stricter policies on both platforms, AI or not.

If you test many creatives per product, these mistakes multiply quickly. Our guide on how many ad creatives to test per product helps you keep volume manageable enough to review every ad properly.

ChecklistChecklist you can copy

CreaXHow CreaX helps (briefly)

CreaX is an all-in-one AI ecommerce creative OS. Paste a product link or upload a product image and, in about 60–120 seconds, it generates UGC-style video ads, studio ad creatives, a publish-ready landing page and a multi-tenant store. It is built for dropshippers, media buyers, affiliates and ecommerce brands that need to test products quickly.

Two things make compliance easier in practice. First, because the ads and the page come from the same product input, the product and the promise start out consistent, which helps with the ad/page mismatch problem. Second, faster generation gives you time back for review. What CreaX does not do is make the legal call for you: you still check every claim, choose the right labels for each platform and follow the rules for your category. Plans start at $4.99 a month for 2,000 credits.

Want to compare tools first? Our roundup of the best AI UGC video ad generators for dropshipping covers other options.

FAQMore questions sellers ask

This guide is published by CreaX. It summarises public Meta and TikTok help pages and ad policies as of October 2026 and is not legal advice. Policies change and vary by region and category, so check the current versions before launching ads.