YouTube has updated the feature for likeness detection. The developers designed the feature to help individuals safeguard their identities from misuse through AI-generated content.
The tool, initially rolled out last year for YouTube creators in the YouTube Partner Program, is now expanding to a pilot for government officials, journalists, and political candidates.
How does the Likeness Detection Tool work?
Just like Content ID, the YouTube likeness detection feature helps determine whether the participant’s image is included in videos produced by AI. Once identified, such as in deepfake videos, participants can view and request removal if it violates YouTube’s privacy policy.
Balancing Safety and Free Expression
YouTube states that “detection is not removal.” There are always exceptions for videos in the public interest, like parody and satire, which YouTube has always allowed in order to promote free expression.
Protecting Participation
To avoid misuse, participants must verify their identities before participating. YouTube claims that the data collected through the process is used only for verification. The tool is not for developing AI content.
Final Thoughts
YouTube intends to extend the service beyond the pilot group in the near future. Meanwhile, the company is actively promoting legislation like the YouTube NO FAKES Act, which grants a federal right of publicity and can be the world’s guide to safeguarding human creativity in the face of AI.