How to label AI content on TikTok, Instagram and YouTube
Quick answer. All three platforms ask the same question, and it is not “did you use AI”. It is whether the result looks like a recording of something real. Realistic generated footage must be disclosed on TikTok, Instagram and YouTube. Animation, stylised visuals and AI used only for scripts or captions generally need no label. The setting is in the post options on each platform, and the duty sits with the account that posted — yours.
The test that actually matters
It is tempting to read these policies as “AI means label it”. That is not what any of them say, and the difference matters in both directions. Label footage you actually filmed and you have made a false statement, which TikTok treats as a violation in its own right. Skip the label on a realistic generated scene and you are the account that published undisclosed synthetic media.
Ask instead: could a viewer scrolling past mistake this for a recording of a real person, a real place or a real event? If yes, disclose. If the video is obviously animated, obviously stylised, or the AI never touched what is on screen, the answer is usually no.
TikTok
TikTok's requirement covers AI-generated or significantly edited content showing realistic scenes or people. You have more than one way to satisfy it: a clear caption, a sticker or a watermark counts, and there is an AI-generated content switch in the post settings which is the unambiguous route.
Two things worth knowing before you upload. First, TikTok sometimes applies the label itself — when you use its own AI effects, or when the file you upload carries Content Credentials from the tool that made it. Once TikTok has applied that label you cannot remove it from the post. Second, TikTok states that turning the setting on does not affect distribution, which removes the usual reason people hesitate.
The reverse is also enforced. Labelling unaltered content as AI-generated is against TikTok's terms and can get the post removed. The label is a factual claim, not insurance.
TikTok is where most AI-made brand content on this platform ends up, so it has a page of its own: AI-made brand content on TikTok covers which shapes are accepted there and which are refused.
Instagram and the rest of Meta
Meta shows an AI info label. Disclosure is required when you post organic content with photorealistic video or realistic-sounding audio that was digitally created or altered, and Meta has said it may apply penalties to accounts that post that without using the disclosure tool.
One detail catches people out: where AI only edited the content rather than generating it, Meta moves the label into the post's menu instead of showing it on the post. So a video of yours may carry the label without it being obvious on the surface. Meta also lets an account describe itself as an AI creator at profile level, which is a standing declaration rather than a per-post one and does not replace the per-post disclosure.
YouTube
YouTube asks an AI use question during upload in Studio. Answer yes and a label goes on the video; on Shorts that label can appear on the player itself rather than underneath.
YouTube is the most explicit of the three about what does not need disclosing, and it is worth knowing because it covers most of what creators actually use AI for:
- Productivity uses — generating a script, brainstorming ideas, automatic captions.
- Unrealistic content, including animation.
- Minor edits such as beauty filters and colour adjustment.
What does need disclosing is realistic content: a real person appearing to say or do something they did not, altered footage of a real event or place, or a realistic scene that never happened. Shorts made with YouTube's own generative tools are disclosed automatically. Repeatedly choosing not to disclose can cost you the content or your place in the Partner Programme.
The label you also need
Every paid post carries a second, separate disclosure: the platform's own paid-partnership label. It answers a different question — who paid for this — and ticking the AI box does nothing for it. On a paid AI clip you need both, and where each control lives is worth reading once properly.
What this means on a campaign
On OVERCLIP, campaigns that accept AI-made content ask you to confirm both labels when you attach the clip. If either confirmation is missing the submission is refused and the reason says which one. That is deliberately stricter than a checkbox you could tick without thinking: the post is already live by the time you submit, so there is no fixing it afterwards without deleting the video and the views with it.
Campaigns that have not opted in do not accept AI-made video at all, and finance, health and gambling briefs never do. Each campaign page states which it is. The full rules are here.
Policies move. The three above were read from each platform's own pages on 26 October 2026; if you are reading this much later, check the platform before relying on the detail.
Read next: where AI-made UGC wins and where filmed still wins, or whether you can actually get paid for AI videos.