More and more AI-generated clips are being uploaded to YouTube without being labeled as such—despite the requirement in place since 2024 to disclose how realistic-looking AI videos were created. Now, Google intends to make improvements by introducing automatic detection; furthermore, the AI label is to be placed more prominently for photorealistic content as well as content that has been significantly altered or generated by AI.
AI labeling to become more visible on YouTube
For regular videos on YouTube, the label will now be displayed directly below the video player and above the description, while for Shorts, it will appear as an overlay directly within the video. (The status of how this will appear for embedded, standard YouTube videos is not mentioned.)
This is intended to allow viewers to recognize at a glance how authentic the images are. For unrealistic, animated, or only slightly altered content, viewers will continue to find this notice in the expanded description.
Starting in May 2026, new internal signals will also be introduced to better automatically detect AI-generated content during upload. If a creator does not disclose that AI was used, but YouTube's systems detect significant use of photorealistic AI, a label will be added automatically in the future.
Should a creator believe that their content was incorrectly classified as AI-generated, they will be able to adjust the status in YouTube Studio, provided the video was not created using YouTube's AI tools, such as Veo or Dream Screen, or if C2PA metadata indicates that the content was created entirely by generative AI. In these cases, the evidence is clear, and the AI label will remain.
It remains to be hoped that the detection will work reliably in practice and that the labeling will contribute to fewer viewers falling for fake videos.
