- Examples that need disclosure and examples that do not
- The AI-use setting at upload and where labels appear
- Why disclosure, copyright and monetisation are separate questions
The short answer
Using AI does not mean every video needs a disclosure. YouTube Help asks you to disclose content that is generated or meaningfully altered with AI in ways viewers could mistake for reality. Go through each place you used AI and compare it with the examples in the help page.
What needs disclosure
YouTube Help lists cases such as making a real person appear to say or do something they did not, altering footage of real events or places, and generating realistic scenes that never happened. Creating music that is central to the video with AI is also included. Deepfakes, synthetic footage made to look like a real person, are a typical case of this idea.
Clearly unrealistic content and minor edits that do not mislead viewers are listed as not needing disclosure.
| How AI was used | How YouTube Help treats it |
|---|---|
| Realistic tornado heading towards a real city | Example needing disclosure |
| Making a real person appear to be arrested | Example needing disclosure |
| AI-generated music | Example needing disclosure |
| Clearly unrealistic footage, such as riding a unicorn | Example not needing it |
| Ideas for scripts, thumbnails or titles | Example not needing it |
| Captions, upscaling, audio repair | Example not needing it |
| Narration using a clone of your own voice | Example not needing it |
Music, voices and images can fall on either side depending on how they are used. As a fictional example, narration in a clone of your own voice is on the "not needed" list, while audio that makes someone else say what they never said follows the "needed" idea. For uses not in the table, do not apply a blanket rule; reread the examples.
Mainly altered or synthetic content that is hard to tell from reality.
Disclosure, whether you may use the material (copyright) and monetisation are separate questions.
The upload setting and labels
According to YouTube Help, when uploading in YouTube Studio you answer the AI-use question in the attributes section with yes if the content meets the requirements and no if it does not. The help page we read was the Android version, and the layout may differ on a computer, on iPhone, and in the English interface.
For photorealistic AI content, a label may appear on the video player; for content that is not photorealistic, or animated, it may appear when the description is expanded. YouTube may add a label automatically, for example to content made with YouTube's own generative AI tools or content containing C2PA provenance metadata.
Disclosure is separate from copyright and monetisation
YouTube Help says disclosing AI content does not restrict the audience and does not affect monetisation eligibility. That does not mean disclosure makes a video monetisable, or that using AI makes it ineligible. Check your right to use the material under the AI tool's terms and copyright, and monetisation under YouTube's monetisation policies, as separate questions. The help page also warns that consistently and deliberately failing to disclose may lead to penalties such as content removal.
Pre-upload checklist and where to check
Video title:
Where AI was used (video / voice / music / images):
Could it pass for a real person, place or event:
Closest example in YouTube Help:
AI-use answer: Yes / No
Checked the terms for the materials used:If you cannot decide, check the latest examples on the YouTube Help disclosure page.
Checklist
- Listed where AI was used, split into video, voice, music and images
- Considered whether it could be mistaken for a real person or event
- Compared it with the examples that need and do not need disclosure
- Answered the AI-use question at upload
- Checked material rights and monetisation separately from disclosure
One thing to check today
List where AI was used in your video and compare each use with YouTube Help's examples of what does and does not need disclosure.