AI-generated ads are allowed on Meta and TikTok. Both platforms treat AI as a technique you may use, and both ask for disclosure in specific cases, mainly when an ad is significantly edited or fully generated, or when an AI-generated person appears.
This guide is for people who run paid social with AI-made creative, especially AI UGC video: what each platform says today, where the risk sits, and a checklist to run through before you launch. Policies change, so each section points to the platform's own page. Read those before you spend, and treat this as general information, not legal advice.
What does Meta say about AI-generated ads?
Meta's published position is about labelling. In its announcement on expanding generative AI transparency for ads products, Meta says it applies an "AI info" label to ads created or significantly edited with its own generative AI creative tools, and that it will automatically detect ads created or edited with third-party AI tools through industry-standard signals and apply the label when it finds them.
Where the label appears depends on the content. Meta states that when its tools result in an AI-generated photorealistic human, the label appears next to the Sponsored label instead of behind the three-dot menu.
Separately, Meta requires disclosure when AI is used to create or alter ads about social issues, elections or politics. A product ad is not in that category, but check the current rules for your own case.
Two practical points follow from this:
- The label is Meta's decision as well as yours. You may not get to choose whether a detected AI ad carries it.
- A label is a notice to viewers. It is not a rejection, and it does not replace Meta's advertising standards on misleading claims.
What does TikTok say about AI-generated ads?
TikTok's advertising policy is more explicit about advertiser action. On its misleading and false content page, TikTok says significantly edited media and AI-generated content is allowed if you apply the AIGC label, or add your own clear disclaimer, caption, watermark or sticker. It says that if undisclosed AI-generated content is found, the ad can be rejected or restricted.
TikTok separates significant edits from minor ones. Content with fully AI-generated images, video or audio counts as significant. Minor tweaks such as adjusting lighting, brightness or colour saturation do not. TikTok also tells advertisers who are unsure to label the content.
Ads Manager has a way to declare an ad as AI-generated, and TikTok describes an AIGC label shown on the ad. Check Ads Manager for the current wording and placement.
How do the AI ad disclosure rules compare on Meta and TikTok?
| Question | Meta | TikTok |
|---|---|---|
| Are AI-generated ads allowed? | Yes | Yes, with the label or a clear disclaimer |
| Who applies the label? | Meta, for its own tools and for AI detected through industry signals | You, through the AIGC label or your own disclaimer |
| What triggers it? | Meta's AI tools, detected third-party AI signals, a photorealistic AI human | Fully AI-generated or significantly edited images, video or audio |
| Minor edits? | Check Meta's current page for when its tools trigger a label | Lighting, brightness, saturation and similar edits do not count |
| Risk if you skip it | Detection and labelling; check the current enforcement terms | Ad can be rejected or restricted |
Read the table as a map to the source pages. Platform wording moves, and your account manager or the help centre is the final authority.
Which AI ads carry the most risk?
The label question is the easy part. The ads that get into trouble are usually the ones that deceive, whether AI made them or not.
| Ad pattern | Why it is risky |
|---|---|
| An AI creator presented as a real customer | It creates a false impression about who is speaking |
| A claim of personal results ("I lost 6 kilos") spoken by a synthetic person | The result never happened to anyone |
| Before-and-after imagery generated rather than photographed | It shows an outcome that was not produced |
| A health, weight or income claim | Heavily restricted regardless of how the ad is made |
| A product shown doing something it cannot do | The generated frame misrepresents the item |
A label cannot cure any of these. Both platforms' honesty rules apply to the claim itself. TikTok's own page says AI content must not be used to mislead people about products, people or facts, so write the script from claims you can support, and do not give a synthetic person a testimonial.
What is the pre-launch checklist for AI UGC ads?
Run this before each launch, per platform.
- Decide whether the ad is fully AI-generated, significantly edited or only lightly touched. Fully generated creators and scenes are the clear case.
- Apply the platform's disclosure. On TikTok, use the AIGC label or a clear disclaimer in the ad. On Meta, expect an AI info label and do not remove provenance data.
- Check every claim in the script. Each statement must be true of your product and supportable.
- Check the creator is not framed as a customer. Describe the person as a presenter or avoid first-person results.
- Check the product frame. Packaging, size, colour and function must match the real item.
- Re-read the current policy pages. Open the Meta and TikTok pages linked above, and Google's ads policy if you run there.
- Keep a record. Save the script, the date and the policy versions you checked, so you can answer a review.
The same list works as a weekly check if you ship AI creative in volume. The creative testing guide covers how to set up that rhythm.
What changes for AI UGC ads disclosure?
AI UGC ads copy a style that viewers read as a real person talking, which is exactly why the disclosure and honesty rules matter more here than for a plain product shot. A synthetic creator saying "I've used this for a month" is a claim of experience. Write scripts around what the product does and what a viewer can see, and avoid scripts that depend on the presenter having had a personal experience. The UGC script template has structures that work with that limit, and the comparison of AI avatars and real creators covers when a real person is the better choice.
Where Klaspr fits
Klaspr is an ad creator for ecommerce that makes UGC video, statics and product shots from a product and a brand kit. The ads it makes are AI-generated, so the disclosure steps above apply to them like any other AI creative. You set the labels and disclaimers in each ad platform, and you are responsible for the claims in the script.
Klaspr does not decide whether a platform will approve an ad, and it cannot read a platform's policy for you. Plans are on the Klaspr pricing page.