YouTube Expands AI Likeness Detection to Celebrities
YouTube revealed plans to broaden its likeness detection system on Tuesday. This tool spots AI-created videos, including deepfakes, that use faces from the entertainment sector. The move aims to shield public figures from unauthorized use of their appearances.
Similar to Content ID System
The likeness detection operates much like YouTube's Content ID. That established feature checks uploaded videos for copyrighted material. Rights holders can then ask for takedowns or claim revenue shares from those videos.
For likeness detection, the focus shifts to fake faces. Creators and notable individuals face issues where scammers exploit their images. Celebrities often discover their likenesses in fraudulent schemes. This technology lets them act against such misuse.
YouTube launched the tool in a limited pilot last year for select creators. It grew this spring to cover politicians, government officials, and journalists. Now it reaches entertainment professionals.
Rollout to Entertainment Industry
Talent agencies, management firms, and the stars they represent can now access the system. YouTube lists backing from key players: CAA, UTA, WME, and Untitled Management. These groups provided input during development.
Participants do not need personal YouTube channels to enroll. The system searches for AI content matching a registered person's facial features. Options include requesting video removal for policy breaches, filing copyright claims, or ignoring the match.
Stay ahead of the AI curve
The most important updates, news, and content — delivered weekly.
No spam. Unsubscribe anytime.
YouTube maintains some flexibility. It allows parody and satire under its guidelines, so not every detection leads to deletion.
YouTube, owned by Google since its 2005 acquisition, handles billions of video uploads yearly. Content ID, introduced in 2007, has managed countless copyright disputes. Agencies like CAA, founded in 1975, and WME represent top Hollywood talent. UTA started in 1991, while Untitled Management focuses on digital influencers.
Future Enhancements and Advocacy
Audio detection will join the features down the line, according to YouTube. The company pushes for wider safeguards too. It backs the NO FAKES Act in Washington, D.C. That bill targets unauthorized AI replicas of voices and images.
No specific figures exist yet on deepfake removals via this tool. YouTube mentioned in March that actions taken remained very small in number.
Deepfakes pose rising threats across media. Politicians faced manipulated videos during elections. Journalists report unauthorized clips spreading misinformation. Celebrities endure scams and explicit fakes without consent. YouTube's step addresses these in video uploads, a platform with over 2 billion monthly users.
The NO FAKES Act, or Nurture Originals, Foster Art, and Keep Entertainment Safe Act, gained bipartisan support. Sponsors aim to create civil remedies for victims of AI likeness theft.