YouTube reveals major change to help you identify AI-generated content – UK Times

YouTube is rolling out more prominent AI disclosure labels and a new detection system designed to automatically identify AI-generated or AI-edited content as it becomes increasingly difficult to distinguish what’s real.

Creators will still be required to disclose when they use AI, but YouTube is adding automated tools that can flag “significant photorealistic AI use,” YouTube announced in a blog post Wednesday. Starting this month, if a creator fails to properly label a video, the platform may apply the disclosure automatically based on its detection systems.

If a video is incorrectly labeled, creators will be able to adjust the disclosure through YouTube Studio. However, YouTube says some AI labels will be permanent, including content created with its own generative AI tools such as Veo or Dream Screen, as well as videos that contain C2PA metadata indicating they were fully generated by AI.

“We’ve heard consistently from our community that they value transparency when it comes to generative AI content,” YouTube’s blog post states. “These changes are designed to balance transparency with creator control.”

The change comes as more and more AI is being used online. It has led some users to complain that distinguishing between real and fake videos online have become harder. That has led to calls for more transparency by media companies to clearly label AI videos.

The platform is also changing where those labels appear. For long-form videos, the AI tag will now sit directly beneath the video player, while on YouTube Shorts, it will show up as a visible overlay during playback.

“The goal here is context at a glance. If it looks real but was made with AI, viewers will know immediately,” Rene Ritchie, YouTube head of editorial and creator liaison, said in a video about the changes.

Ritchie added that the AI labels alone “do not affect how our videos are recommended or whether they can earn money. This is purely about giving viewers the right information at the right time.”

Meanwhile, YouTube is taking a lighter approach to content that is only partially AI-influenced. Videos labeled as “unrealistic, animated or slightly altered,” rather than fully AI-generated, will still get an AI disclosure, but it will appear in the expanded description instead of as a more prominent on-screen label.

The changes follow YouTube’s March move to expand its likeness-detection program to all creators aged 18 and older. That tool is designed to help users detect and manage how AI is used to represent their likeness on the platform.

Once enrolled, creators can receive alerts when YouTube’s systems detect videos that may use synthetic or altered versions of their face. They can then review the content and request removal of unauthorized videos directly through YouTube Studio.