YouTube and Hollywood Agencies Take on Deepfakes
Le brief IA que les pros lisent chaque soir
Les 7 actus IA du jour, décryptées en 5 min. Gratuit.
Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.
Choisis ton rythme
Gratuit · Pas de spam · Désabonnement en 1 clic
YouTube has decided to intensify its fight against deepfakes by expanding access to its AI-based likeness detection technology to talent agencies and Hollywood management companies. Initially reserved for creators, politicians, and journalists, this technology is now available to celebrities through partnerships with renowned agencies such as CAA, UTA, WME, and Untitled Management. This initiative allows actors, musicians, and artists represented by these agencies to benefit from this tool, whether or not they have a YouTube channel.
To use this technology, each participant must first verify their identity. YouTube assures that the data collected during this verification is used solely for this purpose and is not integrated into Google's AI models.
A Content ID for Faces
The detection system implemented by YouTube is inspired by the Content ID model, which is already used for copyright management. However, instead of focusing on audio or video clips, this tool scans newly uploaded videos to detect the faces of registered individuals, including in AI-generated content. When a match is identified, the concerned individual receives an alert and can request the removal of the offending video.
Celebrity deepfakes are often used in fraudulent advertisements or for commercial impersonation. This is why YouTube has chosen to develop this tool in direct collaboration with CAA, UTA, WME, and Untitled Management, to ensure it meets the specific needs of their clients.
Detection Does Not Mean Automatic Removal
While YouTube's initiative is welcomed, it is not without limitations. Detecting a deepfake does not automatically lead to its removal. Once a match is reported, the concerned individual must submit a formal request for YouTube to review the content in question. The platform then evaluates the request based on its own privacy rules. Parody, satire, or political commentary content may remain online despite a removal request. However, for celebrities, whose deepfakes are often linked to scams or commercial counterfeits, YouTube is more inclined to remove them.
YouTube also supports the NO FAKES Act, a proposed federal law in the United States aimed at establishing a right to image protection. Amjad Hanif, YouTube's vice president of creator products, also mentioned that the platform is considering expanding this detection technology in the future to recognizable voices and popular characters.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.