As AI video models become more powerful, YouTube automatically labels videos on behalf of creators, instead of relying solely on creators to label AI videos. The company announced Wednesday that it will apply the label if its internal systems detect that “significant photorealistic AI” has been used.
YouTube also plans to make AI labels more prominent and easier to find on both long-form videos and YouTube Shorts.
The video platform’s AI label has been in use for more than two years, since YouTube updated its AI policy and introduced tools in Creator Studio that require creators to publish videos with AI content that could be mistaken for real people, places, or events. Videos that clearly depicted some sort of animation or imaginative scenario, such as a unicorn flying around in a fantasy world, did not need to be labeled.
The company said its AI labeling policies remain unchanged, but it will take a more active role in policing content on its platform. The move follows Google’s release at last week’s Google I/O developer conference, Gemini Omni, a new family of multimodal AI models that can output high-quality videos that reflect understanding of physics, culture, history, and science.
Starting in May, YouTube will use new internal signals to identify AI-generated content and label it accordingly, the company said. This doesn’t mean creators shouldn’t continue to disclose their use of AI, but failure to do so will result in YouTube labeling their videos.
Creators whose content has been misidentified will be able to update the disclosure status within their YouTube videos, but will not be able to remove those labels if the content was created with YouTube’s own AI tools, such as Veo or Dream Screen, the company said.
If your content includes C2PA metadata, a label will be permanently attached to your video to indicate that it was generated entirely by AI. (OpenAI recently joined Nvidia, Kakao, and Eleven Labs in complying with the C2PA standard.)
The addition of automatic AI detection comes on the heels of YouTube’s expansion of AI deepfake detection, which now allows any adult to scan YouTube for facial matches after initial testing with celebrities, public figures, politicians, and other creators.
YouTube also says it plans to make its AI labels more consistent and prominent.
Previously, a label would appear in the expanded description unless the video touched on more sensitive topics like health or news. In that case, a prominent label will appear directly on the video itself.
The label will now appear directly below the video player, above the description of your long-form video, and will overload directly into your YouTube Shorts.
The company says moving the labels will make them clearer for people who encounter photorealistic, AI-altered, or AI-generated content on the site.
On the other hand, for AI videos that are only slightly modified, animated, or unrealistic (like the aforementioned bouncing unicorn), the label will only appear in the expanded description.
Specifically, YouTube says the AI labels will not affect how videos are recommended or their ability to be monetized.
In addition to regulating AI content, the company has invested in AI for interactive search features, Ask YouTube, YouTube Music playlist generators, AI video summaries, and other generative AI authoring tools.
If you buy through links in our articles, we may earn a small commission. This does not affect editorial independence.
