YouTube is preparing new tools that will allow channel creators and authors to find fake audio and video with them. The company will also allow users to decide how channel content can be used by third parties, including for training AI models.

First of all, as part of the Content ID system, the company is developing a new technology for identifying synthetic singing. It will allow to automatically identify and manage AI-generated content that imitates the voices of real people. Testing of this feature will begin in early 2025.

The second tool is a technology that will allow representatives of all industries, from creators and actors to musicians and athletes, to find and manage AI-generated content that uses their faces.

The company also states that although it uses the content on the platform to train AI, it does so under the terms agreed to by the creators themselves. However, the company also understands that authors want to decide what to do with their videos.

In order to give channel creators more freedom of choice, YouTube is preparing new tools to customize how and what content can be used by third parties, including the ability to train AI.