← Back to tool

YouTube's "Altered or Synthetic Content" Disclosure, Explained

Guide · updated 4 October 2026
Quick answer

YouTube requires creators to self-disclose "realistic" AI-generated or AI-altered content — footage a viewer could mistake for something real — through a toggle in YouTube Studio. As a C2PA steering committee member, YouTube can also carry forward or apply the disclosure automatically by reading Content Credentials (C2PA 2.1+) embedded in the uploaded file, or through its own detection systems, independent of what the creator discloses.

What actually requires disclosure

YouTube's policy is specifically about realism, not AI use in general:

Where the label appears

Most disclosed videos get a label in the expanded description. For sensitive categories — health, news, elections, finance — YouTube shows a more prominent label directly on the video player itself.

C2PA's role specifically

This is the detail most creators miss: if your video (or the image/audio assets inside it) carries valid Content Credentials indicating the entire piece was AI-made, YouTube will carry that disclosure forward automatically — you don't need to manually toggle anything, but you also can't avoid it by simply not disclosing, since the embedded manifest speaks for itself. YouTube may also show an "Info from" attribution naming the signing authority that validated the Content Credentials.

What happens if you don't disclose

For content that should have been disclosed, YouTube can apply the label itself without the creator's input, particularly when the content could confuse or mislead viewers. Consistent non-disclosure carries escalating risk, up to Partner Program suspension in serious or repeated cases.

Checking your assets before upload

If you're compositing AI-generated images or clips into a video and want to know exactly what provenance data they carry before YouTube reads it, AiEraser's View mode decodes the real C2PA manifest on individual image assets, and Remove mode strips it if that's your intent — though for video-level disclosure specifically, remember YouTube's own Studio toggle is the relevant mechanism, separate from an individual image's embedded metadata.

Frequently asked questions

Does using an AI thumbnail image require video-level disclosure?
Disclosure requirements are centered on whether the video content itself is realistic synthetic media a viewer could mistake for real footage — an AI-generated thumbnail alone is a different question from the video's content; check YouTube's current Studio guidance for edge cases like this.
Can I remove a C2PA manifest from my own footage to avoid automatic disclosure?
Technically, stripping C2PA metadata from source assets removes that specific automatic trigger — but YouTube's own detection systems and policies on realistic synthetic content still apply independent of embedded metadata, and misrepresenting AI content as real carries its own risks beyond the labeling system.

Related

Sources: YouTube/Google official blog posts on "altered or synthetic content" disclosure (2024) and "How this content was made" / Content Credentials documentation (support.google.com). Platform policies change; verify current rules directly in YouTube Studio if decision-critical.