YouTube's altered or synthetic content label: when it appears and what it means
YouTube says creators must disclose realistic content made or meaningfully changed with AI tools, and that it adds a label when they do. It also says it may add a label itself, including for content that contains C2PA metadata. A label follows a disclosure or a signal YouTube picks up. We do not read it as a verdict.
What is the label, and who has to disclose?
YouTube's help page, Disclosing use of GenAI content, says creators must disclose AI-generated or altered content that seems realistic. It lists four cases: making a real person appear to say or do something they did not do, altering footage of a real event or place, generating a realistic scene that did not occur, and creating music that is the main focus of the video.
The disclosure is the "AI use" setting in YouTube Studio, under Attributes during upload. The creator answers Yes or No. YouTube says a Yes leads it to add a label for viewers.
YouTube's pages use more than one name for this. The help page's title says "GenAI content". A page on the "Captured with a camera" disclosure, refers to YouTube's "policy on altered and synthetic content". A third page calls the viewer-facing disclosure "Made with AI".
What does YouTube say creators do not have to disclose?
YouTube says realistic content and meaningful changes need disclosure, while non-realistic content and minor edits do not. Its examples of content that is not realistic include a person riding a unicorn through a fantasy world. Its examples of minor edits include beauty filters, color or lighting filters, background blur and captions. It describes these edits as primarily aesthetic.
Where does the label appear?
| Place | What YouTube says |
|---|---|
| Video player | A label may appear for AI-generated music or photorealistic content. |
| Expanded description | A label may appear for non-music AI audio, or for AI content that is non-photorealistic or animated. |
| "How this content was made" section | A "Made with AI" disclosure appears here when a creator discloses, when the creator uses YouTube's generative AI tools, or when valid Content Credentials say the whole video was made with AI. |
YouTube adds that disclosing AI content on its own does not limit a video's audience or its eligibility to earn money. The page we read does not say where labels sit on Shorts, so we do not say.
When does YouTube add a label itself?
The help page says YouTube may apply a label automatically, in the player or the expanded description, for three kinds of content:
- content made with YouTube's own generative AI tools;
- content that contains C2PA metadata;
- content that YouTube's internal systems detect as AI-generated or altered.
YouTube says the creator gets a notification in YouTube Studio. If its systems make an error, creators can in most cases change the disclosure to No. YouTube also says that labels on content made with its own tools, content containing C2PA metadata, and content labeled after manual review cannot be adjusted.
The page adds that creators who consistently choose not to disclose may face a manually applied label or penalties, including removal of content or suspension from the YouTube Partner Program.
What does YouTube say about C2PA metadata?
The two pages word the trigger differently. The disclosure page says "content that contains C2PA metadata". The page on "How this content was made" is narrower: it says YouTube carries forward disclosures from tools and creators with secure Content Credentials (C2PA) 2.1 or higher that indicate the entire video was made with AI. It says such disclosures often name the signing authority, shown as an "Info from" attribution under the disclosure.
The "Captured with a camera" page describes a separate disclosure for video recorded with C2PA-capable tools and not edited afterward. It says a missing disclosure does not mean the audio or visuals were modified.
Google's May 19, 2026 post, Making it easier to understand how content was created and edited, says Google uses C2PA Content Credentials in a growing number of its generative media tools. It names the labels on YouTube that identify AI-generated content as something its work builds on. Our own guide on which platforms show Content Credentials compares other services.
Is having a C2PA record the same as being trusted?
No, and this is a measurement only this site can give. YouTube names the presence of C2PA metadata as one trigger. Presence is not trust. We ran 32 real image files through our checker on October 30, 2026. Nine carry a Content Credentials record, and 23 carry none.
| Headline | Files | Which files |
|---|---|---|
| "Trusted Content Credentials declare AI use" | 1 | A Wikimedia Commons ChatGPT PNG whose certificate names "OpenAI OpCo, LLC" |
| "Content Credentials found; legacy trust-list match" | 5 | Three Adobe Firefly files and two Microsoft Image Creator files |
| "Content Credentials found; signer trust not established" | 3 | Two older OpenAI files whose certificates have expired, and the C2PA test image |
So 1 of the 9 reads as trusted on the current C2PA Trust List. Third parties uploaded the Wikimedia Commons files, which were filed under "Images generated by" categories, so we cannot know how each was handled, and we draw no conclusion about any product. These are still images, not video, and the pages we read do not say which signers YouTube accepts, so we cannot say how YouTube would treat any of these files. The full counts are on our accuracy page.
What this means when you check an image
Our checker reads images only: JPEG, PNG, WebP, AVIF, HEIC and GIF. We tested a one-second MP4 and a one-second MOV on October 30, 2026, and both read "Unable to check this file". We cannot check a YouTube video.
It can still help with a still image, such as a frame you saved or a thumbnail. Open the checker, choose the file, and look for a headline such as "Content Credentials found; signer trust not established". It says the signing certificate names a signer that no trust list we load vouches for, so the name is not verified. It does not say who really signed, or how the image was made.
No headline, and no YouTube label, shows how a picture or video was made. The checker's own line applies: these checks cannot determine whether an image is AI-generated.
Guides on this topic
- Visible AI watermarks is the main page on marks and labels you can see.
- Which platforms label AI images? compares what each service says it labels.
- TikTok's AI-generated label covers the same question for TikTok.
- What LinkedIn's Content Credentials label means covers a label built on signed records.
Frequently asked questions
- Does YouTube label every video made with AI tools?
- The pages we read do not say that. They say creators must disclose realistic content, that it may add labels itself, and that minor edits and non-realistic content do not need a disclosure.
- Does the label limit views or earnings?
- YouTube says disclosing AI content alone does not limit a video's audience or its eligibility to earn money. It separately says repeated non-disclosure can bring penalties.
- Does a missing label mean a video was made without AI tools?
- The pages we read do not say it does. They describe labels that follow a creator's disclosure or something YouTube adds itself, and they say creators who consistently do not disclose may face penalties. The "Captured with a camera" page says only that a missing disclosure of that kind does not mean a video was modified.
- Can your checker check a YouTube video?
- No. It reads still images only. MP4 and MOV files we tested read "Unable to check this file".