YouTube's AI label: when to tick 'AI use' and what it means
The short version
YouTube's AI label tells viewers a video has content meaningfully changed or made with AI. Tick Yes under 'AI use' for realistic fakes and AI music that's the main focus, not for captions or AI scripts.
You’re uploading a video, and YouTube Studio asks about AI use: Yes or No. Captions and AI-written scripts are a No. Realistic fakes are a Yes.
Here’s how the YouTube AI label works, using YouTube’s own examples, as of October 2026.
What YouTube’s AI label means
YouTube’s rules are on a help page called Disclosing use of GenAI content. It says you must disclose when you “use AI to meaningfully alter or generate photorealistic content.” GenAI means generative AI: tools that make images, video, voices or music.
Photorealistic means it looks like real footage. When YouTube announced the rule in 2024, it described content “a viewer could easily mistake for a real person, place, scene, or event.” So ask: could a viewer believe this really happened? If yes, you need the label.
AI music is one extra case: it needs the label when it’s the main focus of the video. YouTube’s page doesn’t say whether AI background music under your voice needs the label. If the music is what viewers came for, pick Yes.
When you must tick ‘AI use’: YouTube’s own examples
In YouTube’s words, you must disclose AI 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’s the main focus of the video
That includes videos only partly made with AI. Here are YouTube’s own examples, which it says aren’t a complete list.
| YouTube’s example (some shortened) | Tick ‘AI use’? | Why |
|---|---|---|
| AI generated music | Yes | Covers AI music that’s the main focus |
| AI generated extra footage of a real place | Yes | Real place |
| Making it appear as if someone gave advice that they did not actually give | Yes | Real person |
| A realistic tornado moving toward a real city that didn’t actually happen | Yes | Fake event, real city |
| Making it look like a real person has been arrested or imprisoned | Yes | Real person |
| Someone riding a unicorn through a fantastical world | No | Not realistic |
| Beauty filters or background blur | No | Minor edit |
| Caption creation, upscaling or audio repair | No | Minor edit |
| Cloning one’s own voice to create voice overs or dubs | No | Your own voice |
Here’s a made-up example. You film a 6-minute vlog in a real beach town, but your beach shot comes out blurry. So you use an AI video tool to make an 8-second realistic shot of that beach.
Before: you pick No. It’s only 8 of 360 seconds, about 2% of the video.
After: you pick Yes. YouTube lists “AI generated extra footage of a real place,” and its own example is a surfer in Maui for a travel video. Partly-AI videos count too.
When you don’t need to
Skip the label for content that isn’t realistic, and for minor edits that are “primarily aesthetic” and don’t mislead viewers. Besides the table, YouTube’s no-list includes:
- AI help with an outline, script, thumbnail, title or infographic
- Idea generation
- Color, lighting and vintage-effect filters
- An AI backdrop that makes a car look like it’s moving
- Gameplay footage from video games
- AI animation inside a fully animated video
Your own cloned voice is fine. A realistic clone of someone else’s voice makes a real person say something they didn’t, so it needs the label. More in is it OK to use an AI voice on YouTube.
Another made-up example. An AI chatbot like ChatGPT outlines your 12-minute tutorial on cleaning up podcast audio. You film yourself teaching it, then make a microphone thumbnail with an AI image tool.
Before: you pick Yes, worried that any AI use counts.
After: you pick No. YouTube lists AI help with an outline or thumbnail as production help that needs no label. Nothing in the video is a fake.
How to tick ‘AI use’ in YouTube Studio
From YouTube’s help page (computer version):
- Go to YouTube Studio (studio.youtube.com) and start uploading your video.
- Find the Attributes section.
- Under AI use, select Yes if the video meets the rules above. If not, select No.
- Fill in the other details and publish.
The Studio phone app has the same setting. To change a video that’s already live, open Studio, select Content and click the video. Change your answer under AI use, then select Save.
If your screen asks about Altered content instead, see the section on older guides below.
What the label does, and doesn’t do
- It tells viewers. The label shows in the expanded description. For photorealistic content, it may also show on the player.
- It doesn’t limit the audience or block earning. YouTube says disclosing “won’t limit a video’s audience or impact its eligibility to earn money.” Whether AI videos can earn at all is a separate policy: can you monetize AI videos on YouTube.
- YouTube can add it for you. It does this for videos made with its AI tools, files with C2PA metadata (hidden data that records how a file was made), and videos its systems detect as AI. You’ll get a notice in Studio.
- You can usually undo a wrong one. Select No under Attributes. That doesn’t work for videos made with YouTube’s tools, C2PA files or labels added after a manual review.
- Skipping it has a cost. Creators who “consistently choose not to disclose” may get a label added, lose videos, or be suspended from the YouTube Partner Program (the program that pays creators).
Older guides call it ‘altered or synthetic content’
In a March 18, 2024 blog post, YouTube called this “altered or synthetic content.” Many older guides still use that phrase. In plain words, “altered” means changed and “synthetic” means made by AI.
As of October 2026, the main help page says “GenAI content,” and the setting is AI use. But YouTube’s upload and edit guides still describe an Altered content setting, under Show more at the bottom of the Details page. It links to the same disclosure page, so answer it the same way.
Why a video you’re watching has an AI label
YouTube’s “How this content was made” help page calls this label Made with AI. It means a video “contains content that has been meaningfully altered or synthetically generated.” It sits in the “How this content was made” part of the expanded description.
It’s there because the creator picked Yes, used YouTube’s AI tools such as DreamScreen, or YouTube added it. It doesn’t mean the whole video is fake, since the rule covers partly-AI videos too.
Common mistakes
- Hunting for an old toggle. Look for AI use under Attributes. If Studio shows Altered content instead, answer that.
- Picking Yes for captions or an AI outline. Both are on YouTube’s no-list.
- Skipping the label for one AI shot of a real place. YouTube lists AI footage of a real place as needing it.
What to do next
- Before your next upload, check each AI shot against the table above. If one is a realistic fake, pick Yes.
- Planning a channel built on AI video? Read can you monetize AI videos on YouTube first.
- Want dated updates on AI rules across apps? See AI news for creators.
Sources
- Disclosing use of GenAI content support.google.com
- Understanding 'How this content was made' disclosures on YouTube support.google.com
- Upload YouTube videos support.google.com
- Edit video settings support.google.com
- How we're helping creators disclose altered or synthetic content blog.youtube
Spotted a mistake? See our corrections policy.