YouTube rolls out AI likeness detection tools for celebrity protection

YouTube is expanding its new “likeness detection” technology, designed to identify AI-generated content such as deepfakes and help protect public figures, the company announced on Tuesday.

The system works similarly to YouTube’s existing Content ID tool, which automatically detects copyrighted material in uploaded videos and allows rights holders to either request removal or claim a share of the revenue. In this case, however, the focus is on detecting simulated or AI-generated faces instead of copyrighted media.

The new likeness detection feature is intended to protect creators and public figures from unauthorised use of their identity, a growing issue as AI-generated scam ads and manipulated videos become more common online.

YouTube first introduced the technology through a pilot program last year, initially rolling it out to a small group of creators. Earlier this year, the company expanded access to include politicians, government officials, and journalists.

Now, YouTube says it is extending availability further into the entertainment sector, including talent agencies, management companies, and the celebrities they represent. The company has received feedback from major agencies, including Creative Artists Agency (CAA), United Talent Agency (UTA), William Morris Endeavour (WME), and Untitled Management.

Importantly, participation in the system does not require users to have a YouTube channel. Instead, enrolled individuals are added to a detection system that scans uploaded videos for AI-generated visual matches of their face.

When a match is found, users can choose to request removal under privacy policy violations, file a copyright removal request, or take no action. YouTube also notes that it will not remove all detected content, particularly where parody or satire is protected under its rules.

The company has also said that future versions of the tool will expand to include audio detection, allowing the identification of cloned or synthetic voices.

Alongside this rollout, YouTube has also backed legislative efforts such as the proposed NO FAKES Act in the United States, which aims to regulate the unauthorised use of AI-generated recreations of voices and likenesses. While YouTube has not disclosed how many removals have been processed using the tool so far, it previously stated in March that the number of takedowns remains “very small.”