AI disclosure for documentary channels: when you have to set “AI use” to Yes
Informational only, not legal advice. YouTube and stock libraries change their pages often; follow the linked official source if it says something different from this guide. Full disclaimer.
Documentary channels use AI in more ways than almost any other format: synthetic narration, upscaled archive footage, AI-generated reconstructions of events no camera recorded, AI music beds, AI-assisted scripts. Some of that has to be disclosed. Most of it does not. The difference comes down to one word in YouTube’s rule: realistic.
Everything below is from YouTube’s help page Disclosing use of GenAI content as it read on 15 October 2026, plus the original March 2024 announcement. Where we interpret, we say so.
The rule in four lines
YouTube requires creators to disclose AI-generated or altered content “that seems realistic.” Specifically, content that:
- makes a real person appear to say or do something they didn’t do;
- alters footage of a real event or place;
- generates a realistic scene that didn’t actually occur;
- creates music that is the main focus of the video.
This applies whether the content is “fully or partially” AI-made. You disclose in YouTube Studio during upload: in the Attributes section, under “AI use”, choose Yes.
What you do not have to disclose
The help page lists examples that need no disclosure. The ones that matter for documentary work:
- “Production assistance, like using generative AI tools to create or improve a video outline, script, thumbnail, title, or infographic”
- “Idea generation” and “Caption creation”
- “Video sharpening, upscaling or repair and voice or audio repair”
- “Color adjustment or lighting filters” and “Special effects filters, like adding background blur or vintage effects”
- “Cloning one’s own voice to create voice overs or dubs”
- Clearly unrealistic content, such as “Using an AI-generated or altered animation of a missile in a fully animated video”
So: AI-assisted scripting, AI-upscaled archive film, colour grading and a cloned version of your own voice all sit on the “no disclosure needed” side.
What a documentary channel does have to disclose
From the help page’s examples of content that needs disclosure:
- “AI generated music”
- “AI generated extra footage of a real place”
- “Showing a realistic depiction of a tornado or other weather events moving toward a real city that didn’t actually happen”
- “Making it appear as if someone gave advice that they did not actually give”
- Realistic depictions of public figures doing things they did not do
Translated to the documentary format, our reading is:
| What you did | Disclose? | Why |
|---|---|---|
| Photoreal AI “reconstruction” of a historical battle or disaster | Yes | A realistic scene of a real event that was not filmed; closest to “AI generated extra footage of a real place” |
| AI-generated realistic shot of a real city “in 1890” | Yes | Realistic footage of a real place that was never recorded |
| AI makes a real historical figure speak or move (animated photo, synthetic voice) | Yes | Makes a real person appear to say or do something |
| Stylised, clearly illustrated or animated AI visuals | Usually no | “Not realistic” examples |
| AI-upscaled or colourised archive footage | Usually no | “Upscaling or repair” and colour adjustment are listed as minor. Our view: if colourisation is heavy enough that viewers could mistake it for original colour film, disclosing costs nothing |
| AI music as the main focus of the video | Yes | Listed explicitly |
| AI music bed under narration | Unclear | The rule says music “that’s the main focus”, but the examples list just says “AI generated music”. If in doubt, disclose |
| Script written with AI help | No | “Production assistance” |
| Your own cloned voice narrating | No | Listed explicitly |
The AI-narrator question
This is the one we get asked most. A licensed synthetic narrator voice (not a clone of any real person) reading your own script is not on either list on the help page.
The 2024 announcement gives “synthetically generating a person’s voice to narrate a video” as an example requiring disclosure, but it sits under the heading “Using the likeness of a realistic person”, which points at imitating a real, identifiable person. Our reading is that a generic stock voice narrating original content is not what the rule targets, while a voice that imitates a real person (a famous narrator, a historical figure) clearly is. That is an interpretation, not YouTube’s wording, and it is exactly the kind of thing that could be clarified later. Check the help page before relying on it.
Disclosure does not cost you money
People avoid the box because they think it hurts reach or revenue. The help page says the opposite: “Disclosing AI content alone won’t limit a video’s audience or impact its eligibility to earn money.”
Not disclosing, on the other hand, has consequences. YouTube says creators “who consistently choose not to disclose” may get a label applied manually, or penalties “including removal of content or suspension from the YouTube Partner Program.”
YouTube may label you anyway
Per the help page, YouTube can apply a label automatically for content made with YouTube’s own AI tools, content that contains C2PA metadata, and content its systems detect as AI-generated or altered. You get a notification in Studio if that happens.
This matters if your AI video tool writes C2PA provenance data into exports: the label can appear even if you selected No. YouTube says that if its systems get it wrong you can usually change the answer under Attributes, but not for content made with YouTube’s AI tools, content with C2PA metadata, or content labelled after manual review.
Where the label shows
For photorealistic content and AI music, the label may appear on the video player itself. For other cases, such as non-music AI audio or non-photorealistic content, it appears in the expanded description.
Disclosure and the monetization policy are separate questions
Ticking the box answers “is this realistic synthetic media?” It does not answer “is this original enough to monetize?” The monetization policy separately lists “AI-generated content made with generic or unoriginal templates giving the impression of mass production” as ineligible (monetization policies). A fully disclosed but templated AI channel can still be demonetized. See reused vs inauthentic content.
A simple routine
- Keep an “AI assets” tab in your license ledger: tool, what it produced, and whether it is realistic.
- At upload, if any row for that video says “realistic depiction of a real person, place or event”, set AI use to Yes.
- If unsure, disclose. It does not affect monetization; failing to disclose can.
Sources (checked 15 October 2026)
- YouTube Help: Disclosing use of GenAI content
- YouTube Blog: How we’re helping creators disclose altered or synthetic content (18 March 2024)
- YouTube Help: YouTube channel monetization policies