YouTube to Integrate Voice and Facial Detection into Likeness Tools

YouTube announced plans to integrate speaking voice detection with facial detection into its likeness detection tool, which was detailed in a blog post by the company's Vice President on September 23. The tool currently scans newly uploaded videos for the faces of creators who have enrolled in the service. Enrollment requires a government-issued ID and a selfie video, and creators consent to voice processing when they enroll.

The company stated that the voice feature aims to improve overall match accuracy for likeness detection. YouTube stores the creator’s legal name, likeness template, and the selfie video for up to three years from the last sign-in. An optional checkbox allows YouTube to use this data to improve its detection models.