What is YouTube's AI content disclosure policy? | AIR Media-Tech
YOU ARE HERE

What Is YouTube's AI Content Disclosure Policy?

Reading time

9 Min

Last updated

26 Sep 2026

What Is YouTube's AI Content Disclosure Policy?
Table of contents
Checklist
22 Steps to Grow from $500 to $10,000 on YouTube.pdf

Since AI has become more prevalent (and much more realistic), YouTube requires creators to disclose when a video contains realistic AI-generated or AI-altered content that a viewer could mistake for real footage. The rule targets deception, so writing a script with AI, generating a thumbnail, or using a generic text-to-speech voice doesn’t require a label.

Realistic synthetic content that could pass as real footage needs the "altered or synthetic content" disclosure. Everyday production assistance is exempt. Skip it when required, and YouTube can apply the label for you.

Question

Answer

What is it?

A transparency label for realistic AI-generated or AI-altered video content

When is it required?

When content could be mistaken for real footage of a real person, place, or event

Where does the label appear?

Beneath the player on long-form and as an overlay on Shorts for photorealistic content; in the expanded description for lighter alterations

What if you skip it?

YouTube can force the label on; consistent non-disclosure risks content removal or YPP suspension

What Is YouTube's Altered or Synthetic Content Disclosure?

The altered or synthetic content disclosure, now labeled "AI use" in YouTube Studio, is a setting creators enable when a video contains realistic AI-generated or AI-altered media a viewer could mistake for real-life footage. YouTube announced the framework on November 14, 2023, and the Studio toggle went live in March 2024. Disclosure doesn’t make otherwise-prohibited content available, because Community Guidelines still apply whether or not something is labeled.

What Triggers the Disclosure Requirement and What Doesn't

Requires disclosure

Example

Realistic synthetic face or voice

An AI-generated person delivering what looks like a real news broadcast

Cloned voice of a real person

Audio altered to make a public figure appear to say something they didn't

Fabricated real event

AI-generated footage that appears to show a real disaster, protest, or accident

Altered footage of a real place

Synthetic scenery making a real location look different than it is

Deepfake of a real person

Video making a real individual appear to say or do something they didn't

AI-generated music

A backing track or song generated with a tool like Suno or Udio

For worked examples by channel type, including AI voiceovers, AI music and dubbing, see AIR's guide to YouTube AI disclosure rules.

Okay, we learned what requires AI disclosure, but what doesn’t?

No disclosure needed

Example

AI-assisted scripts, titles, outlines

Using an AI tool to write copy

AI-generated thumbnails

Still images made with an AI image tool

Generic AI voiceover

Text-to-speech narration in a voice that doesn't imitate any real, recognizable person

Your own cloned voice

A clone of your own voice for voiceovers or dubs of your own videos

Clearly animated or stylized visuals

Cartoons, motion graphics, 3D animation

Unrealistic special effects

Fantasy explosions, sci-fi environments

Standard edits

Filters, color correction, background blur

AI-generated captions or translations

Text-layer tools

The difference is in realism: disclosure applies when a reasonable viewer could believe the content depicts something real. YouTube’s guidance puts it like this: creators aren’t required to ‘disclose content that is clearly unrealistic, animated, includes special effects, or has used generative AI for production assistance.’ More on AI policies on YouTube here.

Using AI voices or visuals and not sure if your channel is at risk?

Every channel mixes AI its own way, and one realistic scene can pull a whole video into the rule. Start with a free channel check, and our specialists leave you a plan of what to review first.

Get my free check →

Where the Label Appears on YouTube

Since May 2026, photorealistic AI content that meets the disclosure threshold gets a label directly beneath the player on long-form videos and as an overlay on Shorts.

Content that's AI-assisted but not photorealistic (stylized or animated AI content, for example) typically only gets the label in the expanded video description, under a heading identifying it as altered or synthetic content.

Viewers see the label as "Made with AI." In the expanded description it sits in a section called "How this content was made," and when the file carries C2PA metadata, an "Info from" line can name who signed it, according to YouTube Help. "Made with AI" doesn't mean a video is fake. It means the creator or YouTube flagged content that was meaningfully altered or generated with AI.

How to Add the Disclosure in YouTube Studio

  1. Upload or edit a video → Details → Attributes section
  2. Under Attributes, find "AI use" and select Yes
  3. Save

The disclosure can also be added after a video is already published:

YouTube Studio → Content → select the video → Details → update the same Attributes toggle

What Happens If You Don't Disclose

YouTube can identify undisclosed synthetic content itself, through its own internal detection systems, C2PA metadata embedded in a file, or manual review, and apply the label without the creator's input.

Since May 2026, YouTube's systems also look specifically for photorealistic AI content and apply the label on their own, as an overlay on Shorts and beneath the player on long-form videos, even when a creator never answered the disclosure question. When that happens, you get a notification in YouTube Studio. If the system labels a video by mistake, in most cases you can change it through the AI disclosure survey under Attributes. Three kinds of labels can't be changed: those on videos made with YouTube's own AI tools, such as Veo or Dream Screen, those on files carrying C2PA metadata, and those applied after a manual review.

Beyond the forced label, YouTube has stated it will look at further enforcement for creators who consistently choose not to disclose, up to content removal or suspension from the YouTube Partner Program for repeat cases.

A note on the policy timeline:

Automatic labeling of undisclosed photorealistic AI content began in May 2026, when YouTube announced the change on its blog on May 27. A separate policy — renamed from "repetitious content" to "inauthentic content" on July 15, 2025, and tightened further in a January 2026 enforcement wave — targets mass-produced and templated content more broadly, AI or not, and affects monetization eligibility rather than disclosure. The two shouldn't be confused.

Does Disclosing Affect Reach or Monetization?

No. YouTube has said directly that adding the disclosure label doesn't affect a video's recommendations, search ranking, reach, or monetization eligibility. The penalty structure applies to not disclosing when required.

This article explains how YouTube's disclosure system works; it isn't legal advice on AI content regulation.

Making AI videos at scale and worried about losing YPP?

The label won't cost you revenue. The inauthentic content policy can. With an AIR channel audit, our specialists and AI agents review your channel and give you a plan ranked by impact.

  • A 10-block report on your channel, channel risks included
  • An action plan ranked by impact
  • A 45-minute live session with a specialist (Channel Growth Blueprint)
  • The experience of a team working with 3,000+ partner channels
Get a plan for my channel →

YouTube
rolled out a drop!
We explained it.

Watch image

Hit our socials,
all the news are there.

FAQ

Do I need to disclose if I used AI to write my script?

No. Using AI for scripts, titles, outlines, or captions is production assistance, not the kind of realistic synthetic media the disclosure policy covers. Disclosure is about content a viewer could mistake for real footage, not about which tools helped make it.

What happens if YouTube labels my video without my input?

YouTube can apply the label itself when its systems, C2PA metadata, or manual review detect undisclosed synthetic content. If the label was applied by mistake, you can update the disclosure status in YouTube Studio; labels on videos made with YouTube's own AI tools, on files carrying C2PA metadata, or applied after a manual review stay. YouTube also sends a notification in Studio when it applies a label itself. Which labels you can still change is covered step by step in AIR's guide.

Does the AI disclosure label affect monetization?

Disclosing itself doesn't; YouTube has confirmed the label alone doesn't change reach, ranking, or monetization eligibility. Failing to disclose when required is what carries risk, up to content removal or YouTube Partner Program suspension for repeat cases.

Is AI content disclosure the same as the "made for kids" setting?

No, though both are mandatory Studio-level disclosures about a video's content. Made for kids exists for COPPA compliance and affects which features and ad types run on a video; the altered or synthetic content disclosure exists to prevent viewers from being deceived about what's real, and doesn't restrict monetization or features on its own.