# AI-Generated Ad Disclosure Rules in 2026 — Cut Marketing

> When do AI-generated ads need a label? What Meta, TikTok, YouTube, Korea's KFTC, the EU and the FTC require as of November 2026, with sources and a checklist.

# AI-Generated Ad Disclosure Rules in 2026

When do AI-generated ads need a label? What Meta, TikTok, YouTube, Korea's KFTC, the EU and the FTC require as of November 2026, with sources and a checklist.

Published 2026-11-10 Last updated 2026-11-16 15 min read [Cut Marketing Team](/editorial)

![AI-generated hand pressing a blank lime tag onto a printed serum photo beside a phone and a clipboard of empty checkboxes](/samples/blog/ai-generated-ad-disclosure-rules-2026.webp)

AI-generated ads are allowed on the major platforms, but most of them now expect a disclosure when the creative is completely AI-generated or realistically altered, and some regulators require a visible mark when an AI-made person recommends a product. As of 10 November 2026: TikTok rejects or restricts ads with undisclosed AI content, YouTube requires creators to disclose realistic synthetic content, Meta adds its own “AI info” label to ads made with its generative tools and says it will begin labeling third-party AI creative it detects, and Korea’s Fair Trade Commission has required “virtual person” marking on endorsements since 1 June 2026.

**In short:**

-   Label by default. The cost of a label is small; the cost of a rejected ad or a deceptive-advertising finding is not.
-   A photorealistic AI person is the highest-risk element everywhere. Product-only shots are lower risk but still count as AI-generated on TikTok.
-   A label does not fix a false claim. An AI presenter cannot give a testimonial.
-   Platform labels and legal duties are separate. Meeting one does not satisfy the other.
-   Rules changed several times in 2026. Check the linked primary sources before a large launch.

This article is a practical summary for information only, not legal advice. Rules differ by country, product category and ad account, so confirm anything that matters with the platform or a qualified adviser. Every rule below links to the primary source we read on 10 November 2026; the TikTok Ads Manager and Google Ads pages were added on 13 November 2026.

## What counts as an AI-generated ad?

An AI-generated ad is an ad whose image, video or audio was created, or significantly changed, by a generative AI system rather than captured by a camera or microphone. Platforms draw the line between “significant” and “minor” in similar ways: generating a scene, a person or a voice is significant; adjusting colour or lighting is minor, and TikTok also lists background removal as minor.

## What does each platform and regulator require?

Each source sets its own trigger and its own duty. TikTok and YouTube ask you to disclose, Meta and Google label some ads themselves, and Korea, the EU and the FTC set legal rules. The table summarizes them, and each section below links the primary text.

| Where | What triggers it | What you do | Consequence stated in the source |
| --- | --- | --- | --- |
| [TikTok ads](https://ads.tiktok.com/help/article/tiktok-ads-policy-misleading-and-false-content) | Completely AI-generated image, video or audio; a subject shown doing or saying something they did not | Apply the AIGC label, or add your own clear disclaimer, caption, watermark or sticker | Undisclosed AI content: “your ad will be rejected or restricted” |
| [YouTube](https://support.google.com/youtube/answer/14328491) | Realistic altered or synthetic content, such as a realistic scene that did not occur | Turn on the altered or synthetic content disclosure when uploading | Consistent non-disclosure: label applied by YouTube, or penalties including content removal |
| [Google Ads](https://support.google.com/google-ads/answer/17140115?hl=en) | Image or video assets generated or edited with AI; election ads with synthetic or digitally altered content | Optional AI label setting per asset; election ads must tick “Altered or synthetic content” | Label in My Ad Center’s “How this ad was made”; on-ad overlays only for campaigns targeting the EU, India, New York, California, South Korea and Vietnam |
| [Meta ads](https://about.fb.com/news/2025/02/gen-ai-transparency-metas-ads-products/) | Ads created or significantly edited with Meta’s generative tools; detection of third-party AI tools announced | Nothing for ordinary ads; self-disclose for social issue, election and political ads | “AI info” label shown in About this ad, or next to “Sponsored” for an AI photorealistic human |
| [Korea (KFTC)](https://www.korea.kr/briefing/pressReleaseView.do?newsId=156764553) | An AI-generated virtual person recommends or endorses a product | Mark “virtual person” at the start of text posts and next to the person in photos and video | KFTC says it will monitor unmarked ads for correction; experience claims that do not match facts can be unfair labeling and advertising |
| [EU (AI Act)](https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai) | Deep fakes and certain other AI-generated content | Label clearly and visibly | Transparency rules apply from August 2026 |
| [United States (FTC)](https://www.ftc.gov/news-events/news/press-releases/2024/08/federal-trade-commission-announces-final-rule-banning-fake-reviews-testimonials) | Endorsements and reviews in advertising | Never present AI-generated reviews or testimonials as real people’s | Civil penalties against knowing violators of the fake-reviews rule |

## Does TikTok require an AI label on ads?

Yes, for significant AI content. TikTok’s advertising policy on misleading and false content (last updated April 2026) allows significantly edited media and AI-generated content if you apply the AIGC label or add “a clear disclaimer, caption, watermark, or sticker of your own.”

What TikTok lists as significant:

-   content that contains images, video or audio that are completely AI-generated;
-   showing the primary subject doing something they did not actually do;
-   making the primary subject say something they did not say, using AI voice cloning.

What it lists as not significant: adjustments to lighting, brightness or colour saturation, removing or modifying backgrounds, and denoising.

The policy’s own advice when unsure is to “use caution and label it accordingly.” The consequence is stated plainly: “If we identify AI-generated content that has not been disclosed, your ad will be rejected or restricted.” Source: [TikTok Advertising Policies, misleading and false content](https://ads.tiktok.com/help/article/tiktok-ads-policy-misleading-and-false-content).

**For product advertisers:** a clip generated from your product photo is completely AI-generated video. Apply the AIGC label when you create the ad, or add a small “AI-generated” caption in the edit. In TikTok Ads Manager the label is a toggle during ad creation, “This ad contains AI-generated content”, for ads that are not Spark Ads; Spark Ads follow TikTok’s Community Guidelines disclosure instead ([Add disclaimers to ads in TikTok Ads Manager](https://ads.tiktok.com/resources/help/article/add-disclaimers-to-ads)).

## What do YouTube and Google Ads require?

YouTube requires disclosure of realistic altered or synthetic content in uploads. Google Ads offers an optional AI label for ad assets and requires disclosure on election ads with synthetic content.

YouTube asks uploaders to disclose content that is realistic and was meaningfully altered or generated. Its help page lists content that makes a real person appear to say or do something they did not, alters footage of a real event or place, or “generates a realistic scene that didn’t actually occur.”

Not required: clearly unrealistic content, minor edits such as beauty filters, and production help such as drafting a script or generating captions. YouTube also states that disclosing AI content alone “won’t limit a video’s audience or impact its eligibility to earn money,” and that creators who consistently choose not to disclose may get a label applied by YouTube or face penalties, including removal of content. Source: [Disclosing use of GenAI content](https://support.google.com/youtube/answer/14328491).

**For product advertisers:** Shorts and most YouTube video ads start as uploads to your channel, where this setting lives. A realistic AI scene of your product in a kitchen that never existed can fit the “realistic scene that didn’t actually occur” description. YouTube’s page does not give a product-shot example either way, so the cautious choice is to turn the disclosure on. The ad itself has its own settings in Google Ads, covered next.

**Google Ads labels.** Starting in July 2026, Google Ads permits text or visual AI labels inside image and video creatives and is gradually rolling out an asset-level AI label setting ([Updates to AI labeling requirements](https://support.google.com/adspolicy/answer/17257106?hl=en)). Labeled ads show the disclosure in the “How this ad was made” section of My Ad Center; overlays on the ad itself appear only for campaigns targeting the EU, India, New York, California, South Korea and Vietnam. Google may also apply labels itself where the law requires it, and says the setting does not guarantee compliance with any regulation ([Use AI content label settings and disclosures](https://support.google.com/google-ads/answer/17140115?hl=en)). Election ads are stricter: advertisers must tick “Altered or synthetic content” for election ads with synthetic or digitally altered content, and must add their own clear and conspicuous disclosure in formats where Google does not generate one ([Google Ads political content policy](https://support.google.com/adspolicy/answer/6014595?hl=en)).

## How does Meta label AI-generated ads?

Mostly automatically, so ordinary advertisers have little to switch on. Its announcement, updated on 1 June 2026, says:

-   ads “created or significantly edited using our generative AI creative features” get an AI info label;
-   the label sits inside “About this ad” in the three-dot menu, but when Meta’s tools add an AI-generated photorealistic human it appears next to the “Sponsored” label;
-   if the tools make no significant edit and include no photorealistic human, no label is applied;
-   Meta “will also begin automatically detecting ads created or edited using third-party AI tools through industry-standard signals” and label those in About this ad too.

Source: [Expanding GenAI transparency for Meta’s ads products](https://about.fb.com/news/2025/02/gen-ai-transparency-metas-ads-products/). Meta’s help page adds that advertisers must themselves disclose digitally created or altered media in ads about social issues, elections or politics, including when third-party AI tools were used: [how AI-generated images in ads are identified and labeled](https://www.meta.com/help/artificial-intelligence/355108217670024/). When we checked, that help page still listed ads made with third-party AI tools among those that do not show AI info, so treat third-party detection as rolling out rather than universal.

**For product advertisers:** there is no AI toggle to forget for an ordinary product ad, but assume the label may appear. Do not strip provenance metadata from files to avoid it, and make sure the ad is one you are comfortable seeing labeled.

## What must Korean ads with an AI virtual person show?

Since 1 June 2026, an ad in which an AI-generated virtual person recommends a product must say that the person is virtual, at the start of text posts and next to the person in photos and video.

Korea’s Fair Trade Commission revised its Review Guidelines on Labeling and Advertising of Recommendations and Endorsements, effective 1 June 2026. The revision adds AI-generated virtual persons as a type of endorser and sets out how to mark them:

-   when a virtual person that is hard to tell from a real one recommends or endorses a product, the ad must disclose that the endorser is a virtual person generated by AI;
-   in text-centred media such as blogs and online communities, put wording such as “\[가상인물 포함\]” (virtual person included) in the title or at the very start of the post;
-   in photos and video, show “가상인물” (virtual person) close to the person, in a colour that stands out from the background, for as long as the person appears;
-   even with the mark, if the virtual person’s recommendation is presented as based on experience and does not match facts that actually happened, it can still be unfair labeling or advertising.

Source: [KFTC press release of 31 May 2026](https://www.korea.kr/briefing/pressReleaseView.do?newsId=156764553) (in Korean).

Two more things to know about Korea:

-   The **AI Basic Act** has been in force since 22 January 2026. Its Article 31 puts transparency duties on AI business operators, meaning those who develop AI or provide products and services built on it: advance notice that a service runs on generative AI, marking of generative outputs, and clear notice for synthetic sound, images or video that are hard to tell from real ones. The duty is written for operators, not for an advertiser who simply uses a tool’s output, but check your own position. Source: [the Act on the national law portal](https://www.law.go.kr/%EB%B2%95%EB%A0%B9/%EC%9D%B8%EA%B3%B5%EC%A7%80%EB%8A%A5%EB%B0%9C%EC%A0%84%EA%B3%BC%EC%8B%A0%EB%A2%B0%EA%B8%B0%EB%B0%98%EC%A1%B0%EC%84%B1%EB%93%B1%EC%97%90%EA%B4%80%ED%95%9C%EA%B8%B0%EB%B3%B8%EB%B2%95) (in Korean).
-   On 10 December 2025 the government announced a plan for an AI-content labeling system under which people who post AI-made photos and video must label them and platforms must manage that labeling ([joint ministry press release, as republished by Seoul city](https://news.seoul.go.kr/economy/?p=569904), in Korean). Whether and when that takes effect should be confirmed against current law before you rely on it either way.

## What do the EU and the United States require?

The EU requires clear labels on deep fakes under the AI Act’s transparency rules from August 2026. We found no general federal AI-label rule for US ads, but the FTC’s rules on endorsements and fake reviews apply, and some states add their own.

-   **EU.** The European Commission states that the AI Act’s transparency rules come into effect in August 2026 and that “certain AI-generated content should be clearly and visibly labelled, namely deep fakes.” Source: [European Commission, AI Act](https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai).
-   **United States.** We did not find a general federal AI-label rule for ads as of the date above, but deception law applies. The FTC’s guidance says “Endorsements must reflect the honest opinions or experiences of the endorser” and “You can’t talk about your experience with a product if you haven’t tried it.” Its rule on fake reviews covers reviews by “someone who does not exist, such as AI-generated fake reviews,” and lets the FTC seek civil penalties against knowing violators. Sources: [FTC Endorsement Guides FAQ](https://www.ftc.gov/business-guidance/resources/ftcs-endorsement-guides-what-people-are-asking), [FTC final rule announcement](https://www.ftc.gov/news-events/news/press-releases/2024/08/federal-trade-commission-announces-final-rule-banning-fake-reviews-testimonials). State law can add to this: New York’s bill S8420-A, signed in December 2025 with an effective date 180 days after it became law, requires a person who produces or creates an advertisement to “conspicuously disclose” a synthetic performer in it ([New York State Senate](https://www.nysenate.gov/legislation/bills/2025/S8420/amendment/A)). Check the states you advertise in.

## Which types of AI creative need a label?

The more realistic the AI element, and the closer it comes to a person vouching for the product, the higher the risk. Real photos with minor edits are low risk; an AI presenter giving a personal review should not run at all.

| Creative | Risk level | What to do |
| --- | --- | --- |
| Real product photo with background removed or lighting fixed | Low | Listed as minor edits by TikTok; YouTube does not ask for disclosure of colour or lighting adjustments |
| Generated product scene or product-only video | Medium | Label on TikTok; disclose on YouTube when the scene looks real; expect Meta’s AI info label |
| AI presenter demonstrating or explaining a product | High | All of the above, plus a visible “virtual person” mark near the presenter for Korean audiences |
| AI presenter giving a personal review or result | Do not run | A label does not cure a made-up testimonial |
| Likeness or voice of a real person without permission | Do not run | Platform policies and laws on likeness, impersonation and deception can all apply |

In Cut Marketing (Cutma), the first two rows are what [Product Video](/tools/create-video-ad) and [Image Generator](/tools/image-generator) produce, and the third is what [UGC Product Video](/tools/ugc-product-video) and [Presenter Video](/tools/lip-sync) produce. We generate the footage, and Finish the Ad in the studio puts your on-screen mark into the video as exact text. Connected Meta, Google Ads and TikTok accounts receive the creative as a paused ad, so you set each platform’s disclosure toggle before you turn it on. Presenter clips do not guarantee exact spoken dialogue, which is one more reason to keep claims in captions you control.

## What should you check before launch?

Confirm the creative type, each platform’s disclosure setting, the Korean marks where they apply, and that no claim rests on an AI person’s experience.

-   [ ]  The creative type is identified: minor edit, generated product scene, AI presenter.
-   [ ]  TikTok: the AIGC label is on, or a clear on-screen disclaimer is in the video.
-   [ ]  YouTube: the altered or synthetic content disclosure is on for realistic scenes.
-   [ ]  Google Ads: AI-generated or AI-edited assets carry the AI label setting where your markets require it, and election ads tick “Altered or synthetic content”.
-   [ ]  Meta: the ad is acceptable with an AI info label, and file metadata is intact.
-   [ ]  Korea: a “가상인물” mark sits near any AI presenter for the whole time they are on screen, and “\[가상인물 포함\]” opens any text post.
-   [ ]  No AI person claims personal use, results or expertise.
-   [ ]  No real person’s face or voice is imitated.
-   [ ]  Every claim, price and offer is true and supportable.
-   [ ]  Paid partnerships and affiliate links carry their own disclosure, separate from the AI label.
-   [ ]  The date you last checked each platform policy is written in the campaign notes.

For the wider question of when an AI presenter is appropriate at all, see [AI UGC ads vs real creators](/blog/ai-ugc-ads-vs-real-creators). For platform sizes and safe zones, see [TikTok, Reels and Shorts ad specs for 2026](/blog/tiktok-reels-shorts-ad-specs-2026).

## What are the common mistakes?

Most of them come from treating a platform label as the whole job.

-   **Treating the platform label as legal cover.** A TikTok AIGC label does not make a false claim acceptable.
-   **Labeling the video but not the text.** In Korea a blog post or community post that features a virtual person needs the mark at the start of the text as well.
-   **Removing metadata to dodge a label.** It works against the disclosure the policies ask for.
-   **Assuming product-only means exempt.** On TikTok, completely AI-generated video is significant whether or not a person appears.
-   **Copying last year’s rules.** Meta updated its ad labeling announcement on 1 June 2026, the same day Korea’s guideline took effect, and the EU’s transparency rules began to apply in August 2026.

## FAQ

### Do I have to label AI-generated ads?

Often, yes. TikTok requires a label or your own clear disclaimer on ads with completely AI-generated or significantly AI-edited content. YouTube requires disclosure of realistic altered or synthetic content. Meta applies its own AI info label to ads made with its generative AI tools and says it will begin labeling third-party AI creative it detects. Korea’s KFTC requires virtual-person endorsers to be marked.

### Does Meta require advertisers to disclose AI-generated ads?

For ordinary product ads Meta applies the AI info label itself, in About this ad, when ads are created or significantly edited with its generative tools, and it says it will begin doing the same for third-party AI tools it detects through industry-standard signals. Advertiser self-disclosure is required for social issue, election and political ads.

### What happens if I do not label an AI-generated ad on TikTok?

TikTok’s advertising policy says that if it identifies AI-generated content that has not been disclosed, the ad will be rejected or restricted. You can disclose with TikTok’s AIGC label or with your own clear disclaimer, caption, watermark or sticker.

### Do AI product shots without people need a label?

It depends on the platform. TikTok counts completely AI-generated images or video as significant, so label them. YouTube asks for disclosure when a realistic scene that did not actually occur is generated. TikTok lists lighting adjustments and background removal as minor edits, and YouTube does not ask for disclosure of colour or lighting adjustments.

### What does Korea require for AI presenters in ads?

Since 1 June 2026 the KFTC’s endorsement review guidelines require ads in which an AI-generated virtual person recommends a product to disclose that the person is virtual: at the title or start of text posts, and next to the virtual person for as long as they appear in photos and videos.

### Can an AI presenter say they used my product?

No. An AI presenter has no experience to report. The FTC’s Endorsement Guides say endorsements must reflect the honest opinions or experiences of the endorser, and Korea’s guidelines say experience-based claims by a virtual person that do not match real facts can be unfair advertising even when labeled.

### Does Cut Marketing label my ads for me?

Not automatically. You can add the on-screen disclosure line yourself as exact text in Finish the Ad, in the studio. Cut Marketing can also upload approved creatives to your connected Meta, Google Ads and TikTok ad accounts and create the ads paused, but each platform’s AI disclosure setting and the claims in the ad stay your responsibility. Check them before you turn the ad on in the platform’s ads manager.

## Make the creative, then label it properly

Sign in with Google and [open Product Video in the studio](https://app.cutmarketing.app/studio/create-video-ad?ui_locale=en) to generate product footage from your own photo. An 8-second 1080p Veo 3.1 Fast clip starts from 32 credits ($0.32). Then add captions, your disclosure line and your call to action in Finish the Ad, export 9:16, 1:1 and 4:5 in one go, and upload to your connected ad account as a paused ad. See the [AI video ad generator](/ai-video-ad-generator) page for what the studio makes and [pricing](/pricing) for credit costs.

## Keep reading

[Can You Use AI Images and Videos in Ads? Copyright Rules 2026](/blog/ai-generated-ads-copyright-commercial-use): Yes, you can usually use AI images and videos in ads, but you may not own a copyright in them. What US, Korean and EU law and tool terms say, plus a checklist.

[Meta vs Google vs TikTok Ads for Small Business](/blog/meta-vs-google-vs-tiktok-ads-small-business): Which ad platform a small business should start with in 2026, how Meta, Google and TikTok compare, official minimum budgets and a first-month budget split framework.

[How to Find Competitor Ads on Meta and TikTok (2026 Guide)](/blog/find-competitor-ads-meta-tiktok): Find competitor ads on Facebook, Instagram and TikTok with free public tools, read days running and versions honestly, and log what is worth testing yourself.

## Make the next ad in minutes.

Start with a product image or a short brief. The studio shows the credit cost before you generate.

[Open studio](https://app.cutmarketing.app/?ui_locale=en)
