YouTube expands AI deepfake detection tool to celebrities and public figures

YouTube is rolling out its likeness detection tool to celebrities as part of a broader effort to give public figures control over AI-generated impersonations on the platform. The tool, which scans uploaded content for AI-manipulated facial likenesses, was first tested with creators last fall and expanded to politicians and journalists in March. The expansion to celebrities removes a previous barrier: enrolled figures no longer need a YouTube account to participate.

To use the system, participants must provide government-issued identification and a selfie video. The tool then monitors the platform for matching synthetic content and surfaces matches to the enrolled figure, who can request removal. However, YouTube evaluates each takedown request against its privacy policy, and content deemed parody, satire, or otherwise protected may not be deleted.

YouTube's approach mirrors its existing Content ID system for copyright enforcement, with one key difference: Content ID allows rights holders to monetize infringing videos and claim revenue. The likeness detection tool does not yet offer this option, though industry signals suggest it may be coming. YouTube recently announced a separate feature enabling creators to generate AI clones of their own likenesses, and talent agency CAA maintains biometric databases that entertainers can use to authorize commercial uses.

The entertainment industry's framing of deepfakes has begun to shift. Rather than viewing synthetic media purely as a threat, some talent managers and celebrities are exploring how AI impersonations could deepen fan engagement or generate licensing revenue. High-profile cases like TikTok creator Khaby Lame's attempt to commercialize his likeness rights illustrate this trend, though such deals have faced execution hurdles.