YouTube today announced that it will introduce automatic AI detection and apply automatic AI labeling to videos that “use significant photorealistic AI.”
Even with the new automatic labeling system, creators will still be required to manually disclose when realistic AI is used in their videos. Creators who believe their content has been incorrectly identified as AI-generated can update their disclosure status in YouTube Studio, but disclosures for content created with YouTube’s AI tools like Veo and Dream Screen, or with C2PA metadata that indicates fully AI-generated, are permanent.
YouTube is also improving its labeling of AI-generated content, making it more clear that videos contain “photorealistic and meaningful AI-modified or generated content.”
The AI label appears just below the video player and above the description for long-form videos with AI content. Additionally, for short videos, labels appear as an overlay on the video. The updated label applies to content that is realistic and potentially deceptive, such as photos, rather than unrealistic content.
Disclosures of “unrealistic, animated, or slightly altered” content will continue to be incorporated into the video’s expanded description.
Separately, YouTube has added a new customizable content feed that users can create based on their interests, mood, and favorite topics. Users enter a custom prompt with what they want to see, and a content feed is generated.
Custom content feeds have been in testing since November and are now rolling out to logged-in viewers on the YouTube mobile app and desktop in the US. To use this feature, YouTube search and watch history must be turned on.
