TikTok rules for AI-generated content in 2026
TikTok requires a label on realistic AI people and scenes. Here is every rule, linked to TikTok's own pages.
TikTok allows AI-generated content but requires creators to label any AI content that shows realistic-looking people or scenes, using the "AI-generated content" setting before posting. TikTok can also label posts itself from C2PA Content Credentials. Unlabeled realistic AI content can be removed, restricted or labeled by TikTok. Some content stays banned even with a label, such as AI likenesses of anyone under 18.
Make your first videoTikTok lets you post AI-generated videos and photos, but any AI content that shows a realistic-looking person or scene needs the "AI-generated content" label, and some AI content is banned even when labeled. TikTok can also add the label on its own when a file carries C2PA Content Credentials. Meta reads similar file signals on Instagram and also requires creator disclosure, as covered in how Instagram's AI info label works. If you are about to post AI dance clips, read this before you follow how to make AI dance videos of your AI model.
Proof
We did not run a test on TikTok's systems. The proof is this table of rules from TikTok's own newsroom, help, Creator Academy and Community Guidelines pages, each linked, with the date where the page shows one.
| Date | What TikTok said | Source |
|---|---|---|
| September 2023 | TikTok introduced a label creators can apply to content completely generated or significantly edited by AI, and said it would test an automatic "AI-generated" label | TikTok newsroom |
| May 2024 | TikTok became the first video-sharing platform to implement C2PA Content Credentials, to auto-label AI content uploaded from other platforms that carries the metadata, and said it would attach Content Credentials to content made on TikTok | TikTok newsroom |
| Newsroom post | TikTok is testing invisible watermarks on AI content made with its own tools, such as AI Editor Pro, and on uploads with C2PA Content Credentials, plus a control in Manage topics to see more or less AI content in For You | TikTok newsroom |
| Current page | Creators must label AI-generated or significantly edited content that shows realistic-looking scenes or people; unlabeled content can be removed, restricted or labeled by TikTok | TikTok, AI-generated content policy |
| Current page | No AI content with the likeness of anyone under 18, or of an adult private person without permission; no AI content falsely showing public figures endorsing something or being bullied | TikTok, Integrity and Authenticity |
| Current page | The label sits under More options before posting as "AI-generated content"; TikTok may also label automatically through its AI effects and Content Credentials | TikTok Help Center |
| Current page | Small edits such as colour correction, reframing or cropping need no label; the label does not affect distribution if the content follows the guidelines | TikTok Creator Academy |
This page states only the rules shown on TikTok's current policy pages; read the full Community Guidelines before you post.
What the label rule covers
The test TikTok uses is realism. If a viewer could take the person or the scene for real footage, it needs the label. That catches almost everything an AI model account posts: a generated woman dancing in a kitchen, a talking clip, a face swap, a voice that sounds human.
What falls outside: clearly stylised work such as anime, and ordinary edits like cropping, reframing and colour correction. TikTok's Creator Academy names those directly.
The label has two sources. You can set it yourself before posting. TikTok can also set it when it finds C2PA Content Credentials in the upload, or when the content was made with one of its own AI effects. A file that went through a re-save or an editor that drops metadata gives TikTok nothing to find, which is why the creator label still carries most of the weight. The standard behind those credentials is published by the C2PA, and our guide to the Instagram AI info label shows how Meta reads the same signals.
What stays banned with a label
A label is not a permit. TikTok's Integrity and Authenticity rules ban some AI content outright:
- AI content showing the likeness of anyone under 18.
- AI content using an adult private person's likeness without their permission.
- AI content that falsely shows a public figure endorsing something, or being bullied or degraded.
- AI content that misleads people on matters of public importance, such as fake crises or fake authoritative sources.
For a persona account this means the face has to be a generated person, not someone you know, and never a teen look. When you design one, the face design chapter of the Academy is the place to start, and creating an AI influencer step by step covers the rest of the build.
Method: posting AI video to TikTok
- Make the clip in the AI video generator, or remake a trending format with the Recreate a video flow from the viral clip chapter.
- Check the content against the banned list above before you think about the label.
- Label it. In TikTok, turn on the AI-generated content setting under More options. If you schedule through PixelPig, the studio does this for videos.
- Post natively to TikTok instead of cross-posting, the same advice the Academy gives for every platform.
On step 3, here is exactly what PixelPig's scheduler does, from our code. When you post a video to TikTok from the Content calendar, the request sets video_made_with_ai: true, and the post window tells you "PixelPig marks this video as AI-generated content on TikTok." For photo posts the studio does not set the flag, because TikTok rejects it on direct photo posts from some account types, so the window says to label AI content in TikTok yourself. That logic lives in supabase/functions/social-post/index.ts and src/components/social/PostComposer.tsx. The full posting flow is in the scheduling chapter of the Academy.
TikTok and Instagram compared
| Rule | TikTok | Instagram (Meta) |
|---|---|---|
| What must be labeled | AI content with realistic-looking scenes or people | Photorealistic AI video and realistic-sounding audio |
| Creator tool | AI-generated content setting under More options | Meta's AI disclosure tool in the posting flow |
| Automatic signals | C2PA Content Credentials, TikTok AI effects, invisible watermarks on its own tools | C2PA and IPTC metadata, invisible watermarks from some companies |
| If you skip it | Removed, restricted or labeled by TikTok | Penalties can apply |
| Effect on reach | TikTok says none if the content follows the guidelines | No official statement found |
Common mistakes
Labeling the caption instead of the post. A hashtag such as #ai helps viewers, but TikTok's rule refers to its label setting. Turn the setting on.
Planning to fix the label later. Set it at upload, so the video is labeled from its first view.
Assuming a photo post is covered by the scheduler. PixelPig labels videos on TikTok, not photo posts. Add the label in TikTok for photo carousels.
Reading "no automatic label" as "not detected". TikTok only auto-labels when it finds a signal. The duty to label realistic AI content is yours either way.
Using a real person's face. Even with a label, a private person's likeness without permission is banned.
Bottom line
TikTok allows AI content and asks one thing of creators: label anything that looks real. Use the AI-generated content setting before you post, keep minors, private people and fake endorsements out of it, and set the label at upload. If you post videos through PixelPig's scheduler, it sends the AI flag for you; check the label on the live post, and label photo posts in TikTok.
Questions
- Does TikTok allow AI-generated content?
- Yes. TikTok allows AI content as long as realistic AI people and scenes are labeled and the content does not break the Community Guidelines. Some uses are banned even with a label.
- How do I add the AI-generated label on TikTok?
- Before you post, open More options and turn on the AI-generated content setting. TikTok's help page describes it there. The post then shows that the creator labeled it as AI-generated.
- What happens if I do not label AI content on TikTok?
- TikTok says unlabeled realistic AI content can be removed, restricted or labeled by its team, depending on the harm it could cause.
- Does the AI label hurt views on TikTok?
- TikTok's Creator Academy says the label does not affect distribution as long as the content follows the Community Guidelines.
- Can I remove the AI-generated label after posting?
- Decide before you publish. TikTok's Creator Academy page on the label covers what can change after posting, so check it there rather than planning to fix the label later.
- Does TikTok detect AI content automatically?
- Partly. TikTok reads C2PA Content Credentials on uploads and labels content made with its own AI effects. It has also announced invisible watermarks on content made with its AI tools and on uploads that carry Content Credentials.
- Do I need to label AI anime or cartoon videos on TikTok?
- TikTok's rule targets realistic-looking scenes and people. Its Creator Academy says clearly stylised work, such as anime, and small edits like cropping or colour correction do not need the label.