-
YouTube has announced a major update to its content transparency policies: the platform will now automatically detect and label videos that contain significant photorealistic AI use. The new system, revealed on Wednesday, May 27, 2026, applies automatic AI labels even when creators have not manually disclosed the use of artificial intelligence.
According to YouTube, the automatic detection focuses on content that is photorealistic and meaningfully AI altered or generated. This means videos that could realistically fool viewers into thinking they are authentic will receive a prominent label.
For content that is unrealistic, animated, or only slightly altered, the disclosure will remain tucked away in the expanded description of the video.
Even with the new automatic detection, YouTube expects creators to continue manually disclosing when realistic AI is used. If a creator believes their content was incorrectly flagged as AI-generated, they can update the disclosure status in YouTube Studio.
However, disclosures are permanent for content created using YouTube's own AI tools like Veo and Dream Screen, or for content that contains C2PA metadata indicating fully generative AI. This ensures transparency for viewers watching AI-generated videos on the platform.
Separately, YouTube introduced a customizable content feed that lets users shape their viewing experience based on interests, moods, or preferred topics. Viewers can type a custom prompt describing what they want to see, and YouTube generates a tailored feed.
The custom feed feature has been in testing since November, but it is now rolling out to signed-in viewers in the United States on the YouTube mobile app and desktop. To use the feature, users must have YouTube search and watch history turned on.
Comment