YouTube introduces new tool to detect deepfake content

AI has dominated discussions throughout this year; however, a troubling aspect has emerged alongside its achievements.

The prevalence of AI-related scams and deepfakes has significantly increased. Deepfake videos featuring celebrities have become alarmingly frequent, raising concerns as these impersonations have grown increasingly realistic, making distinguishing them from genuine content difficult.

In response to this issue, YouTube is intensifying its efforts to safeguard creators and celebrities from the exploitation of their identities. To this end, the platform has formed a partnership with the Creative Artists Agency (CAA).

Together, they intend to introduce sophisticated tools designed to detect AI-generated content.

These tools will empower creators to identify videos that replicate their facial features, vocal characteristics, or other elements of their identity through artificial intelligence.

This advancement will facilitate the process of requesting the removal of such unauthorized content. YouTube plans to launch this initiative next year, initially focusing on celebrities and athletes, followed by a broader rollout to prominent creators, influencers, and other professionals on the platform.

Through this initiative, the streaming giant seeks to address the growing concerns surrounding impersonation and the unauthorized use of AI-generated representations. YouTube first revealed its plans to develop tools for managing AI-generated portrayals of creators in September.

Additionally, YouTube has collaborated with CAA, which represents numerous global stars, to create a system known as CAAVault. This technology maintains comprehensive digital records of its clients’ faces, voices, and likenesses, establishing a database that can be utilized to detect unauthorized usage.

Moreover, to combat the misuse of AI in music, YouTube is developing a new “synthetic-singing identification technology.” This feature aims to identify and eliminate AI-generated content that mimics artists’ vocal performances, a growing concern for music labels. Lastly, YouTube now mandates that creators label AI-generated content.