TikTok requires brands to disclose any post made or significantly edited with AI that shows realistic-looking people or scenes, using its AI-generated content label or a clear caption, sticker or watermark of their own, and it may remove, restrict or label posts that skip it (TikTok Community Guidelines, effective September 24, 2026). TikTok says the label itself does not reduce how widely an eligible post is shown, but some AI content is banned even when labeled, and ads and TikTok Shop carry stricter rules. In the US, FTC endorsement and fake review rules apply on top, and since June 2026 New York requires ads to disclose AI-generated performers.
Key takeaways
- Disclose AI-made or heavily AI-edited posts that show realistic people or scenes, with TikTok’s label or your own clear caption, sticker or watermark (TikTok Community Guidelines). Small edits, anime-style art and generic text-to-speech are exempt (TikTok Creator Academy).
- You cannot edit or remove the label after posting (TikTok says “at this time”), and TikTok says it does not hurt reach for eligible content (TikTok Creator Academy).
- TikTok also auto-labels using C2PA Content Credentials and its own detection models, and says its labeling efforts, creator labels included, had labeled over 1.3 billion videos by November 2025 (TikTok Newsroom).
- Banned even with a label: AI likenesses of minors, private people used without permission, fake public figure endorsements, and content that misleads about real events.
- Ads and TikTok Shop are stricter than organic posts.
What does TikTok require brands to label as AI-generated?
TikTok’s Community Guidelines (effective September 24, 2026) require creators, brand accounts included, to disclose AI-generated or significantly edited content that shows realistic-looking scenes or people. TikTok’s AIGC label is one way to disclose; a clear caption, sticker or watermark of your own also meets the Community Guidelines. Undisclosed content “may be removed, restricted, or labeled by our team, depending on the harm it could cause” (TikTok Community Guidelines).
TikTok defines AI-generated content as any image, video or audio made or changed by AI, but only some of it needs the label, as of October 2026:
| What the post contains | Label needed? |
|---|---|
| AI-made people or scenes that look real | Yes |
| Your real product in a photoreal AI-generated room or scene | Yes |
| A face replaced with someone else’s | Yes |
| AI making someone appear to say something they didn’t | Yes |
| A person, object or background added or removed in a misleading way | Yes |
| AI audio that mimics a real person’s voice | Yes |
| Color correction, reframing or cropping | No |
| Anime or other artistic styles | No |
| Generic text-to-speech that is not a known person’s voice | No |
| Your real product cut out onto a plain background, not misleading | No |
Source: TikTok Creator Academy, last updated September 25, 2026, and TikTok Advertising Policies for background edits. See also our AI-generated content label glossary entry.
Slideshows count: a photo post of photoreal AI people using your product needs the label. And when unsure, label. TikTok’s ad policy says “it’s best to use caution and label it accordingly” (TikTok Advertising Policies).
What AI content is banned on TikTok, even with a label?
TikTok’s Community Guidelines list AI content that is not allowed even with a label (TikTok Community Guidelines):
- The likeness of private figures used without consent. TikTok counts everyone under 18 as a private figure, and its Safety Center says it does not allow AI likenesses of under-18s.
- Likenesses made to bully or harass, and sexualized or victimizing depictions.
- A public figure shown supporting products or taking stances they never took.
- Content that misleads about matters of public importance, such as fake news sources or crisis events.
For brands, the endorsement rule bites first: an AI clip of a celebrity “recommending” your product is out, label or not. And if you turn a real employee or customer into an AI spokesperson, get their consent first.
How do you turn on TikTok’s AI-generated content label?
In the TikTok app, the setting sits in the post options (TikTok Creator Academy):
- Tap the Add post button, then record or upload your video, or capture your photo for a photo post.
- For a video, tap the Continue check mark, then Next, then More options. For a photo post, open Post settings.
- Turn on the AI-generated content setting, then post.
Viewers then see a label below your caption.
If you post through a scheduling or content tool, check whether it sets the label. TikTok’s Content Posting API has an AI-generated content field that tags the post “Creator labeled as AI-generated” (TikTok for Developers) when a tool publishes directly. A post that arrives in your TikTok inbox as a draft needs you to turn the setting on when you post it.
How does TikTok label AI content automatically?
TikTok layers three methods (TikTok Newsroom, November 2025):
- Creator labels from the setting above.
- TikTok’s own detection models.
- C2PA Content Credentials. Content Credentials are a technical standard from the C2PA (Coalition for Content Provenance and Authenticity) that attaches metadata to a file describing how it was made (TikTok Safety Center). In May 2024 TikTok became the first video sharing platform to read Content Credentials and auto-label AI content made on certain other platforms (TikTok Newsroom, May 2024).
TikTok also auto-labels content made with its own AI effects (TikTok Safety Center), and in November 2025 said it would start adding invisible watermarks to AI content made with TikTok tools such as AI Editor Pro and to uploads carrying Content Credentials (TikTok Newsroom).
What this means for a brand: if your AI generator writes Content Credentials into its files, TikTok may label the post for you, and you cannot remove a label the platform applies. If no label appears, you still have to disclose realistic AI content yourself (TikTok Shop Academy).
Will TikTok penalize AI-generated content from my brand?
Not for being AI, and not for being labeled. What TikTok has published, as of October 2026:
- TikTok says the label does not cut reach. “Applying the label will not have an impact on organic engagement with your content, provided that your content isn’t violative and meets our For You feed eligibility standards” (TikTok Creator Academy).
- TikTok Shop says the same. Self-labeled and auto-labeled content is not penalized “solely” for the label (TikTok Shop Academy).
- Skipping disclosure is the risk. Undisclosed realistic AI content can be removed, restricted or labeled by TikTok (TikTok Community Guidelines).
- Copy-paste volume is a separate risk. Content that “includes unoriginal or reused material without anything new” is ineligible for the For You feed (Community Guidelines).
In November 2025 TikTok said it would test a Manage Topics control that lets viewers see more or less AI content in their For You feed. And viewers react to the label: a USC Marshall study of more than 1.1 million TikTok posts from 8,650 creators, plus eight preregistered experiments, found posts with AI disclosures got roughly 7% to 8% fewer likes at the same view count, because the label weakens viewers’ connection to the creator, partly by lowering perceived effort. Disclosures that signaled more creator effort softened the drop; simply naming which part was AI did not (Journal of Consumer Research).
Label honestly, and use the caption to show the human work behind the post (“AI visuals, recipe developed and tested by our team”), since signaling effort is what softened the drop in the study.
What are the rules for AI in TikTok ads and TikTok Shop?
| Where it runs | What must be disclosed | What happens if you don’t | Source |
|---|---|---|---|
| Organic posts | AI or significantly edited content showing realistic people or scenes | Removal, restriction or a label added by TikTok | Community Guidelines |
| Ads, including Spark Ads | Completely AI-generated media, or real footage changed so a person does or says something they didn’t | Ad rejected or restricted | Advertising Policies, Ads Help Center |
| TikTok Shop (US) | Content fully generated or significantly edited by AI | Removal or restricted visibility; repeated or severe cases can lead to account bans | Shop Academy |
Ads. As with organic posts, TikTok accepts its AIGC label or “a clear disclaimer, caption, watermark, or sticker of your own” (TikTok Advertising Policies), and Ads Manager lists AI-generated, synthetic or manipulated media as a mandatory disclaimer case (TikTok Ads Help Center). Minor AI edits such as lighting changes, background changes and denoising do not count (TikTok Advertising Policies). Unlike the organic rule, the ad rule on completely AI-generated media has no “realistic-looking” qualifier. Boosting an organic AI post? Read how to turn organic TikToks into Spark Ads and apply the ad rules first.
TikTok Shop. The US policy dated September 2, 2026 bans using AI to change a product’s size, color, material, quantity, bundle contents, features or performance, and names AI-made “doctors” and expert personas endorsing health products as a key enforcement focus (TikTok Shop Academy). Supplement and skincare sellers should read it, along with our notes on TikTok content for supplement brands.
What do FTC and US state rules mean for AI content on TikTok?
TikTok’s label is about how content was made. The FTC cares whether an ad deceives people. This is a summary, not legal advice:
- Disclose that it is an ad. TikTok requires the content disclosure setting whenever you promote a brand, product or service, including your own business. Your own promotions get a “Promotional content” label and posts for another business get “Paid partnership” (TikTok Help Center).
- Disclose creator relationships inside the video. When you pay creators, the FTC says the disclosure belongs in the video, not only in the description, and that a platform’s disclosure tool may not be enough on its own (FTC, Disclosures 101).
- You answer for your creators. The FTC tells advertisers that “your company is ultimately responsible for what others do on your behalf” (FTC Endorsement Guides FAQ).
- No fake customers. The FTC’s 2024 rule on reviews and testimonials bans testimonials that misrepresent that they come from someone who does not exist, “such as AI-generated fake reviews” (FTC). The FTC’s 2023 Endorsement Guides update also covers virtual influencers.
- New York requires AI performer disclosure in ads. Since June 2026, an advertiser who knows an ad includes a synthetic performer (an AI-generated person who is not a real, identifiable performer) must disclose it conspicuously; penalties are $1,000 for a first violation and $5,000 after (NY Senate, S8420A).
In practice, the risky pattern is an AI person presented as a real customer: “I’ve used this for three months and my skin cleared up.” An AI explainer or an accurate product demo does not pretend to be anyone’s experience.
What should a brand check before posting AI content on TikTok?
- Does it show realistic-looking people or scenes made or heavily changed by AI? Turn on the AI-generated content label, the simplest way to disclose.
- Does it use a real person’s face or voice? Get their permission, never use a minor’s likeness, and never show a public figure endorsing you.
- Does the product look, work and perform the way it does in real life? If AI changed any of that, redo it.
- Is the post promoting your business? Turn on the content disclosure setting, and if a creator posts it for you, disclose the paid relationship in the video itself.
- Does anyone on screen speak as a customer? If they are not a real customer with that experience, rewrite it.
- Will you run it as an ad? Check the ad disclaimer rules first.
How does Relay handle AI labels and disclosures?
Relay (relaystudio.ai) makes TikTok slideshows and videos for your brand, you swipe to approve, and Relay sends each approved post to your TikTok on its day. It lands in your TikTok inbox, where you add any finishing touches and tap Post. Relay does not set the AI label or the content disclosure for you: when you post the draft in TikTok, turn on the AI-generated content setting if the post uses realistic AI imagery, and the content disclosure toggle where it applies.
Relay does not replace a legal review: in a regulated category, have someone check health, financial or before-and-after claims before you approve, and leave Autopilot off so every draft gets that check.
On Scale, Relay’s creator network lets creators post your approved drafts from their own accounts, which brings the FTC creator disclosure rules into play.
When is Relay the wrong tool for AI content?
- You want realistic AI spokespeople. Relay makes slideshows and videos in your brand’s voice, based on formats getting views in your niche. If the plan is AI avatars talking to camera, a tool built for that, such as TikTok’s Symphony Creative Studio with its AI avatar videos, will fit better, and those videos need the label either way. See Relay vs TikTok Symphony.
- You need Instagram or YouTube publishing. Relay’s self-serve product works with TikTok only today.
- You want creative for paid campaigns. Self-serve Relay is built for organic posts, starting with your own TikTok. Relay’s managed UGC ads service is the separate route for ad creative.
The short version
Turn on the label for realistic AI, and never let AI change how your product looks or performs. For a month of TikTok posts made for you and sent to your TikTok inbox, see how Relay makes AI UGC videos or start at relaystudio.ai.