YouTube is launching a new tool to detect videos that use artificial intelligence to replace faces. The purpose of this feature is to combat the spread of unwanted deepfake materials and ensure their rapid removal from the platform. This is reported by Engadget.
So far, only members of the YouTube Partner Program have access to the tool. The system identifies cases where AI changes a person’s face, but so far it does not detect voice changes without the owner’s consent. To use the protection, the user must confirm their identity by providing a copy of a government document and a short video selfie. Additionally, you must upload original footage for verification.
The system works in a similar way to Content ID, which is used to detect copyrighted audio. It scans uploaded videos, finds possible matches, and allows the face owner to review the results to flag violations for removal.
The launch of the tool comes amid growing concerns about the proliferation of video-generated technologies, particularly with the emergence of new models such as OpenAI’s Sora 2.
