JAKARTA - YouTube has finally launched a new likeness detection technology that is ready to hunt down deepfake videos on its platform. After months of testing, this feature is now available to creators around the world to protect their faces and voices from abuse in AI-made content.
This step comes amid the increasing trend of AI videos that are increasingly difficult to distinguish from original recordings. YouTube hopes this similarity detection technology can become a human version of Content ID, not for copyrighted music or videos, but for the face and voice creators.
This feature allows creators to identify and request the removal of videos that mimic their appearance or voice without permission. Selected creators have started receiving invited emails to join. The registration process is quite tight: creators must enter the Likeness tab on YouTube Studio, approve data processing, then verify identity by scanning the QR code, upload official ID photos, and record short selfie videos.
SEE ALSO:
After verification, creators can review videos marked by the YouTube algorithm. They can choose to delete, file copyright claims, or archive the video.
YouTube's representative to TechCrunch confirmed that this is only the initial stage of the launch, and will be extended to more creators in the coming months. The main goal: preventing facial abuse and creator voices, especially in fake advertising or disinformation campaigns.
Interestingly, YouTube also gives creators the full freedom to leave the program at any time. If they choose to stop, the system will automatically stop scanning videos that show their facial resemblance within 24 hours.
This move makes YouTube one of the first major platforms to actively fight deepfakes with a digital identity-based approach. The digital world is getting more and more wild, but at least now there are eyes of algorithms that also protect the faces of its creators.
The English, Chinese, Japanese, Arabic, and French versions are automatically generated by the AI. So there may still be inaccuracies in translating, please always see Indonesian as our main language. (system supported by DigitalSiber.id)