Main points
- YouTube will allow celebrities to track and request removal of deepfake videos, but the final decision will remain with the platform, and there is no provision for automatic blocking.
- The feature is already being tested among bloggers, journalists, and politicians, and now it is planned to be made available to all famous people who confirm their identity and record a video selfie.
Deepfakes on YouTube will start to disappear / Depositphotos
YouTube is expanding its tools to combat deepfakes. The platform will allow celebrities to track videos with their artificially created image and submit requests to have them removed. However, this process will not be automatic, and the final decision will remain with the service.
The developers announced the changes in their approach to combating deepfakes on the YouTube blog.
What is this system and how will it work?
First of all, it is necessary to explain what we are talking about deepfakes – videos where the face or image of famous people is replaced with the help of AI algorithms. Now celebrities themselves will receive a tool that will allow them to find such videos and influence their fate.
The analysis system is based on similarity recognition technology . It searches for content that uses a specific person's face and notifies program participants of any matches found. Public figures can then request the video to be removed .
At the same time, each such case will be considered individually – automatic blocking is not provided, and this is an important aspect of the system's operation to avoid abuse and false activation.
Important note! Not all deepfakes will be subject to removal. For example, videos that are parodic or satirical in nature may remain on the platform. This means that YouTube is trying to balance privacy protection with creative freedom.
When will everything work?
The function was tested in the fall among bloggers, and in the spring journalists and politicians were added to it. Now they plan to open access to all famous people – even those who do not have their own channel. To join the program, you need to confirm your identity and record a video selfie. In this case, the system focuses on the face, and not on the voice or other characteristics.
The technological foundation is the Content ID tool, which was previously used primarily for copyright protection and content monetization. Now its functionality has been expanded to combat fake videos.
The new AI avatar industry
In parallel, YouTube is moving in the other direction – the development of AI content. Earlier, the company announced the ability to create your own digital avatars for videos. And the industry has already begun to commercialize such tools. For example, the Creative Artists Agency is creating a database of biometric data of its clients for the use of their images in commercial projects.
This trend is confirmed by the case of Khaby Lame – a popular TikToker has already sold the rights to use his own image to advertise products on the Internet, as Fortune reported.
As for the attitude towards deepfakes among celebrities, it still remains ambiguous. Some of them are against the use of their image without permission, while others see it as new opportunities for interaction with the audience and even additional earnings. It seems that the industry is gradually moving towards a model where the digital image becomes a separate asset that can be successfully monetized.