YouTube Expands AI Portrait Detection Tool to Entertainment Industry Celebrities
YouTube is extending its newly launched "likeness detection" technology to the entertainment industry to help celebrities and their teams identify and remove AI-generated content and deepfake videos on the platform, as announced on its official website on Tuesday.
This tool operates similarly to YouTube's existing Content ID system, which automatically detects copyrighted material when users upload videos, allowing copyright holders to request takedowns or participate in ad revenue sharing. Likeness detection focuses on identifying simulated faces, aiming to protect creators and other public figures from having their identities exploited without permission – such as the increasingly common issue of celebrity likenesses being used in scam advertisements.
YouTube initially launched this technology as a pilot project with a limited number of creators last year, gradually expanding its coverage. This spring, YouTube extended its application to include groups such as politicians, government officials, and journalists, to help them combat the misuse of AI deepfake content. Now, the company has announced that the tool will further extend to participants in the entertainment industry, including agencies, artist management companies, and the celebrities they represent. Several major Hollywood agencies – including CAA, UTA, WME, and Untitled Management – have participated in providing feedback and refining the new tool.
It is worth noting that using the likeness detection tool does not require the artist to have their own YouTube channel. In the background, the system scans the platform for AI-generated content to detect footage matching the facial information of registered participants. Once a match is found, relevant users can choose to initiate a removal request according to YouTube's privacy policy, submit a copyright takedown request, or choose to take no action. However, the platform also emphasizes that not all related videos will be removed, as its rules still allow for a certain range of parody and satirical content.
YouTube says the technology will also be expanded to support audio dimensions in the future, meaning it will eventually be able to identify and address unauthorized AI-synthesized voices and other content. Alongside product evolution, YouTube is also advocating for more systematic protection measures at the federal level, supporting the "NO FAKES Act" currently being discussed in the U.S. Congress. The bill aims to regulate the unauthorized recreation of individuals' voices and visual likenesses using artificial intelligence technology, providing clearer legal boundaries for public figures and ordinary users.
Currently, YouTube has not disclosed how many AI deepfake content removal requests have been processed by the tool to date. However, the company stated in March that the number of takedowns triggered by the likeness detection tool remains "very small." As AI-generated technology penetrates deeper into the entertainment and advertising industries, an increasing number of public figures – from creators and politicians to film stars – are relying on platform-level technical means to address the risk of their likenesses and voices being misused.