Compliance problems
Flagged as AI-generated? What the labels mean and how to fix it
Short answer
A video gets an AI label in one of three ways: you turned the disclosure on, the platform's own AI tool made part of it, or the file carried content credentials or other signals the platform detects. Realistic AI content has to be disclosed on YouTube, TikTok and Instagram, so the fix is the right setting at upload, not hiding the source. Labels from the platform's tools or from content credentials cannot be removed, but a wrong disclosure can usually be switched off in the post's settings.
Key takeaways
- Realistic AI-generated or meaningfully AI-altered content must be disclosed on YouTube, TikTok and Instagram.
- YouTube and TikTok say automatic labels can come from their own AI tools and from C2PA content credentials.
- A label you applied yourself can be changed; one applied automatically from credentials or the platform's tools cannot.
- Both YouTube and TikTok state that a correct disclosure does not limit distribution or monetization on its own.
- Do not strip metadata to dodge a label: undisclosed realistic AI content is a policy violation.
Three ways a label gets onto a video
| Route | Who applies it | Can you change it? |
|---|---|---|
| You switch on the disclosure | You | Yes, on YouTube in most cases |
| Platform's own AI tool or effect was used | The platform | No |
| Content credentials or detection in the file | The platform | Not on YouTube or TikTok once applied automatically |
This is why the same video can be labeled on one platform and not another: each looks at different signals.
What each platform says needs disclosure
| Platform | Disclose | Not required |
|---|---|---|
| YouTube | A real person made to say or do something they did not; altered footage of a real event or place; a realistic scene that did not happen; music that is the main focus | Non-realistic or minor edits |
| TikTok | Realistic AI images, audio and video; primary subjects shown doing or saying what they did not; a face swap that makes the person unrecognizable | Content made only with TikTok effects is labeled automatically |
| Photorealistic video or realistic-sounding audio that was generated or altered | Images are not required to be labeled but may be labeled if detected |
Find your case
| What you see | What it usually means | Do this |
|---|---|---|
| A label appeared and you never turned anything on | The platform applied it automatically (its own AI tool, content credentials or detection) | Check which route applied it (step 1) |
| You used an AI voice, face swap or generated scene | Realistic AI content has to be disclosed on every platform | Turn the disclosure setting on (step 2) |
| You only used AI for captions, noise removal or color | Minor edits are not what the rules target | Leave the disclosure off and read the platform's examples (step 3) |
| Label cannot be changed in the app | Labels from the platform's own tools or from content credentials are locked | Re-export without the AI-made part, or accept the label (step 4) |
| Reach dropped after the label | Could be a coincidence; the platforms say a correct disclosure does not limit reach | Compare against your last five posts (step 5) |
Fix it step by step
Find out why the label is there
On YouTube, Studio sends a notification when a label is applied automatically. On TikTok, an auto label says AI-generated; one you chose says creator labeled as AI-generated. On Instagram, the label can come from Meta's tools or from detected signals in the file.
Disclose realistic AI content when you upload
Turn on the platform's disclosure setting before posting: Attributes → AI use in YouTube Studio, More options → AI-generated content in TikTok, Add AI label on the Instagram publication screen.
Decide whether your edit needs disclosure
The rules target content that looks real: a real person made to say or do something they did not, altered footage of a real event, a generated realistic scene. Captions, filters, noise reduction and color work are minor edits.
Change a wrong disclosure, if the platform allows it
On YouTube you can switch the AI disclosure to No in Attributes in most cases. Content made with YouTube's AI tools, carrying C2PA metadata, or labeled after review cannot be changed. TikTok says an auto label cannot be removed once applied.
Measure reach against your own baseline
Compare the labeled post with your last five uploads of the same format before concluding the label cost you views. YouTube says disclosure does not limit the audience or affect eligibility to earn.
Do not remove credentials to avoid a label
Older advice says to strip C2PA or other metadata so the platform cannot see it. That does not change what the video is: realistic AI content still has to be disclosed, and YouTube says creators who consistently do not disclose can face penalties. If a label is wrong, correct the disclosure setting; if it is right, the label is the platform working as designed.
A pre-posting checklist for AI-assisted video
- List what the AI did: voice, face, background, a whole scene, captions or cleanup.
- If any realistic person, event or place was generated or altered, turn the disclosure on.
- Check whether a platform effect was used; those are labeled automatically.
- After posting, open the video as a viewer and confirm the label reads the way you expect.
Frequently asked questions
Do I have to label AI content on every platform?
Each platform requires disclosure for realistic AI-generated or meaningfully altered content: YouTube for photorealistic changes, TikTok for realistic images, audio and video, Instagram for photorealistic video and realistic-sounding audio.
Can I remove an AI label?
Only one you applied yourself, and only where the platform lets you. YouTube lets you change the disclosure to No in most cases; labels from the platform's AI tools, C2PA credentials or manual review stay. TikTok says an auto label cannot be removed.
What are content credentials (C2PA)?
Metadata that records how a file was made, attached by some generators and editors. YouTube and TikTok both say they may auto-label content that carries them.
Will an AI label hurt my reach?
YouTube states that disclosing does not limit a video's audience or monetization, and TikTok says the setting does not affect distribution if the video follows its guidelines.
What happens if I never disclose?
YouTube says creators who consistently do not disclose may get a label applied manually or penalties, including removal of content or suspension from the Partner Program. TikTok says unlabeled realistic AI content may be removed.
Is using AI for subtitles or noise removal AI content?
Minor edits and non-realistic effects are not what the disclosure rules are aimed at; the question is whether the result could pass as real footage of something that did not happen.
Sources
- YouTube Help: Disclosing use of GenAI content (official, checked 2026-10-06)
- TikTok Help: About AI-generated content (official, checked 2026-10-06)
- Instagram Help: Label AI content on Instagram (official, checked 2026-10-06)
Reviewed 2026-10-06. Steps not attributed to a source are general video practice or computed above.