YouTube will automatically label some AI made videos

In short: YouTube says it will automatically add labels to videos when it detects realistic AI was used, and it will place those labels where viewers can see them more easily.

What happened

YouTube announced it will no longer rely only on creators to disclose when they used AI in a video. Instead, YouTube says its own systems will add a label when it detects “significant photorealistic AI,” meaning AI that can make something look real, like a real person or event.

The company says the underlying policy is not changing. Creators are still expected to disclose AI use, but YouTube will now label videos even if a creator forgets or chooses not to.

YouTube is also changing where labels appear. For regular, longer videos, the label will show directly below the video player and above the description. For YouTube Shorts, the label will appear on top of the video (like a sticker you can see while watching).

YouTube said creators can update the disclosure status if a video was labeled by mistake. However, if the video was made with YouTube’s own AI tools, such as Veo or Dream Screen, creators will not be able to remove the label.

In some cases, labels can be “locked in” when a video includes C2PA metadata. Metadata is extra hidden information attached to a file, like an ingredient label on packaged food. C2PA is a standard that helps show whether content was AI generated.

Why it matters

More videos can now be made to look real even when they are not. Clearer and more automatic labels can help viewers understand what they are seeing, especially for videos that could be mistaken for real news or real people.

Source: TechCrunch AI