2026-10-31

Do You Have to Disclose AI on YouTube? The Altered Content Label Explained

Every time you upload a video, YouTube Studio asks a question about altered or synthetic content. For faceless creators who use AI images, AI voices or AI help with scripts, it is one of the most confusing boxes on the screen. Tick it when you don't need to and you worry viewers will think less of the video. Skip it when you should have ticked it and you break a rule.

The good news is that YouTube's own help page answers most cases with examples. This post walks through them and applies them to typical faceless videos. It is not legal advice, and the rules are updated from time to time, so read the official page before you rely on any detail here.

Official source: Disclosing use of altered or synthetic content on YouTube Help (checked 31 October 2026). YouTube announced the upload-time disclosure in March 2024 on its official blog.

The rule in one sentence

YouTube's own words: "We require creators to disclose AI-generated or altered content that seems realistic." In practice, that means content a viewer could mistake for a real person, place or event, plus one special case: AI-generated music that is the main focus of the video.

The key word is realistic. The rule is about not misleading viewers into thinking something real happened. It is not a general "did you use AI?" question.

What needs disclosure

YouTube lists four kinds of generative-AI content that need the box ticked:

Making a real person appear to say or do something they didn't. Including realistic synthetic voices of a real, identifiable person, or making it look as if someone gave advice they never gave.
Altering footage of a real event or place. YouTube's example is AI-generated extra footage of a real place in a travel video.
Generating a realistic scene that didn't actually happen. YouTube's example is a realistic tornado moving toward a real city.
Creating music that's the main focus of the video. If an AI song is the point of the video, disclose it.

What doesn't need disclosure

YouTube also lists things that do not need the label. The main groups:

Clearly unrealistic content. Someone riding a unicorn through a fantasy world, a green-screen float through space, an AI-made animation in a fully animated video.
Minor edits and effects. Colour correction, lighting filters, background blur, vintage effects, beauty filters, sharpening, upscaling, audio repair.
Production help. Using AI to create or improve an outline, script, title, thumbnail or infographic, idea generation, and caption creation.
Your own cloned voice. YouTube lists "cloning one's own voice to create voice overs or dubs" as not needing disclosure.

So if you used AI to help draft your script, and nothing on screen is realistic and misleading, the script help on its own doesn't trigger the label.

Applying it to faceless videos

Here is how this plays out for the kinds of videos faceless channels typically make. These are our reading of YouTube's examples, not official rulings. When in doubt, disclosing is the safer choice: YouTube says disclosure alone doesn't limit reach or monetisation.

What's in your video	Likely needs disclosure?	Why
AI helped outline or edit the script	No	Production assistance
Auto-generated or AI-assisted captions	No	Production assistance
Flat 2D illustrations, cartoon or clay-style AI images	Usually no	Clearly unrealistic
Painterly or engraving-style AI images of a historical event	Usually no, if clearly an illustration	Not mistakable for a real photo
Photorealistic AI "photos" of a real historical event	Likely yes	A viewer could take them for real archival photos
Realistic AI image of a real, identifiable person	Yes	Realistic likeness of a real person
A generic AI narrator voice reading your script	Usually no, if it doesn't imitate a real person	YouTube's page has no general rule for stock AI narrator voices; it lists cloning your own voice as not needing disclosure, and synthetic voices of real people as needing it
AI-generated background music under narration	Usually no	YouTube's music example is music that is the main focus of the video; check the page if music plays a bigger role
An AI clone of a real person's voice	Yes (and you need their permission)	Synthetic voice of a real person
AI "news footage" of a disaster in a real place	Yes	Realistic depiction of an event that didn't happen that way
Your own diagrams or maps	No	Not synthetic media

The row that catches most faceless channels is the photorealistic historical image. Many history and mystery channels use AI to create "archival-style" photographs of real events. If the image looks like a real photograph of a real event, a viewer can reasonably believe it is one. That is exactly what the rule is about.

How to disclose

In YouTube Studio (on a computer or the mobile app), the question appears while you upload. As of 31 October 2026 the help page describes it as an "AI use" setting in the Attributes section of the video details: select Yes or No. You can also change the answer later by editing the video's details. The exact screen layout changes over time; YouTube's help page has the current steps.

When you answer yes, YouTube adds a label. For AI-generated music or photorealistic content, a label may also appear on the video player itself. For non-photorealistic or animated AI content, and for AI audio that isn't music, the label may appear in the expanded description.

YouTube can also add a label on its own: for content made with YouTube's own AI tools, content carrying C2PA metadata (an industry standard some AI tools and cameras attach to files), or content its systems detect as AI-generated or altered. You get a notification in Studio. In most cases you can change it by answering "No" in the disclosure survey, but not for content made with YouTube's AI tools, content with C2PA metadata, or content labelled after a manual review.

Does disclosure hurt your video?

This is the fear behind most of the confusion. YouTube's help page answers it directly: "Disclosing AI content alone won't limit a video's audience or impact its eligibility to earn money." The risk is the opposite. YouTube says creators who consistently fail to disclose may get a label added for them, or face penalties up to content removal or suspension from the Partner Program. And viewers who feel tricked leave angry comments and stop trusting the channel.

Good habits beyond the checkbox

The checkbox is the minimum. These habits make a faceless channel more trustworthy, and they are easy:

Label illustrations on screen

If you use a realistic illustration for a real event, a small on-screen note such as "AI illustration" or "Reconstruction, not a photograph" removes any doubt. It costs nothing and viewers appreciate it.

Separate real archival material from illustrations

If your video mixes real archival photos (for example, from a public-domain collection) with AI illustrations, make the difference visible. Use a different frame or caption style for each, or keep AI images in a clearly different style (painterly, engraved, flat). Our post on keeping AI images consistent shows how to pick a style block that reads as illustration.

Avoid real faces

Don't generate images of real, identifiable people. It triggers disclosure, it can raise privacy and likeness problems, and YouTube has a separate process that lets people request removal of AI content that simulates their face or voice. Show people from behind, in silhouette, or focus on objects and places.

Don't use a real person's voice

Use your own voice, a narrator you hired, or a licensed AI voice from a tool whose licence allows commercial use on YouTube. Never clone or imitate a real person's voice without their written permission.

Mention your process in the description

A line such as "Narration: AI voice. Illustrations: AI-generated, not photographs. Sources listed below." is honest and short. Viewers who care can see exactly what they are watching.

Disclosure and monetisation are different rules

Ticking the disclosure box is a transparency rule. Whether a video can earn money is a separate question, decided by YouTube's monetisation policies, especially the rules on reused and inauthentic content. A fully disclosed AI video can still be refused monetisation if it is mass-produced; an undisclosed one can still be original. You need to get both right. We explain the second one in YouTube's inauthentic content policy, explained.

Quick checklist before every upload

Is anything on screen realistic and synthetic or meaningfully altered?
Could a viewer believe it shows a real person, place or event as it really was?
Is there any real person's face or voice that is synthetic? (If so: disclose, and make sure you have permission.)
Is an AI-generated song the main focus of the video?
Have I labelled illustrations on screen where they could be mistaken for photos?
Does my description say what is AI-made and list my sources?
Have I read YouTube's current help page for any new examples?

If you answer yes to either of the first two questions, or to the AI-song question, tick the box.

Plan the look before you generate

The easiest way to avoid disclosure headaches is to choose a visual style that is clearly illustration from the start. The Faceless YouTube Starter Pack includes six tested style blocks, from ink-and-engraving to flat 2D, with example prompts for each.