YouTube Opens Its Deepfake Defense Arsenal to Hollywood

YouTube is now offering its proprietary deepfake detection technology to a broader roster of public figures. The platform, after initial tests with political figures, is extending access to actors, musicians, and athletes concerned about unauthorized AI-generated videos that mimic their appearance.

The tool functions like a digital likeness scanner. A celebrity or their representative can submit reference material, and the system combs through uploads to identify potential matches. Flagged videos are presented for review, and the individual can then request removal, even without a YouTube account. The company emphasizes that not every match will result in a takedown; parodies or satirical works may stay under existing guidelines.

This move responds to growing tension in entertainment circles as AI video generation becomes more accessible. While major studios have targeted AI developers, the sheer volume of user-generated content presents a persistent challenge. YouTube's approach aims to provide a direct line of defense on its own service.

During earlier testing, participants reportedly removed only a small fraction of flagged content, primarily targeting harmful or defamatory material. Mary Ellen Coe, YouTube's chief business officer, described the current phase as building a "foundational layer of responsibility and protection." She acknowledged a future where rights holders might monetize such AI content but stated that is not the present objective. The immediate goal is to offer control, shifting some enforcement burden from the platform to the individuals most affected.

Source: CNET