YouTube is taking a more proactive stance on content transparency by implementing an automated system to label videos containing significant photorealistic AI. While the platform previously relied on creators to self-disclose the use of artificial intelligence, new internal detection systems will now identify and tag such content automatically. This shift ensures that viewers are clearly informed when they are watching media that may depict realistic people, places, or events generated by AI models.

The updated policy, which begins rolling out this month, aims to make these disclosures more prominent across the platform. For long-form content, labels will now appear directly beneath the video player, while YouTube Shorts will feature an overlay label. These changes are designed to improve visibility, particularly for content that could be mistaken for authentic footage. In cases where content is identified as fully AI-generated through C2PA metadata, the labels will remain permanently attached to the video.

Despite these new automated measures, YouTube maintains that its core policies regarding AI disclosure remain unchanged. Creators are still encouraged to label their own content, though the platform will intervene if a creator fails to do so. If a video is incorrectly flagged, creators retain the ability to update the disclosure status, unless the content was produced using YouTube’s proprietary AI tools like Veo or Dream Screen. The company has clarified that these labels are intended for transparency and will not negatively impact a video’s monetization potential or its performance in recommendation algorithms.