YouTube now recognizes AI videos itself – even without self-disclosure from the creators

Anyone who has had enough of “AI slop” should be happy: YouTube is moving the labels for AI-generated content to a more visible place and is also introducing automatic recognition. The video platform can now label videos even without the uploader’s self-disclosure. YouTube announced this in a blog posting.

Where the label appears

Previously, the AI ​​notice below the video only appeared in the expanded description area. Viewers first had to look there specifically to see him. That’s changing now. For normal horizontal videos, the label appears directly below the player. And: With shorts, YouTube displays them as an overlay directly in the video.

YouTube distinguishes between AI-generated videos that are obviously AI-generated for users and AI clips that could be misleading due to their realistic appearance. YouTube only implements the prominent placements for AI videos that appear photorealistic or have been significantly changed by AI and this can hardly be seen, the company writes. For animated AI videos that users can easily recognize, the labels remain in the old position in the opened description text.

YouTube recognizes AI videos independently

Previously, YouTube relied on voluntary self-disclosure from uploaders to label AI videos. That wasn’t particularly reliable because there were no means of sanction. Now YouTube recognizes AI clips independently using an algorithm. To do this, the platform checks, among other things, the metadata of the videos. If AI markers according to the C2PA standard are included, YouTube automatically sets a label. C2PA stands for Coalition for Content Provenance and Authenticity – an open standard that anchors content provenance information directly in the media file.

How accurate the self-recognition is remains to be seen. Incidentally, creators can dispute incorrect labeling in YouTube Studio. However, if content was created using Google’s own tools such as Veo or AI markers are included in the metadata, such AI labeling is irrefutable.

Gradual introduction from May 2026

YouTube will gradually implement the new system from May 2026. The existing disclosure obligation remains unaffected. Anyone who does not declare their use of AI must expect that YouTube will set the label itself. Since spring 2024, creators on YouTube have had to disclose photorealistic AI content.

According to YouTube, AI labeling has no influence on recommendation algorithms and advertising. The company justifies the changes with feedback from the community that has been collected since the disclosure requirement was introduced in 2024. Accordingly, users want to know what AI is, and how many click away has apparently not been checked.