YouTube is enhancing its AI video labeling system to make it more prominent and automated, starting this month. The platform will use "new internal signals" to flag videos with "significant photorealistic AI use," according to the company. Previously, creators had to manually disclose if they used AI tools, but now YouTube will apply labels automatically.
The new labels will appear more prominently on standard videos and YouTube Shorts, with a visible tag below the video or as an overlay on Shorts. Creators can appeal if they believe their videos were incorrectly labeled, but not if the tags are based on C2PA metadata or watermarked Google tools like Veo. The labels are permanent and will be placed in areas where they are more easily visible, such as directly below the video and above the description box.
Google has not specified if the label is clickable, but it appears to be. The system targets "photorealistic and meaningfully AI altered or generated content," meaning some AI-created videos may still use the older, less visible disclosure method.
Source: arstechnica
