YouTube has partnered with Creative Artists Agency to develop technology that will help celebrities identify and remove deepfake videos created by AI to exploit their images. YouTube announced the tech in September and has now gained CAA’s support in the form of “critical feedback to help us build our detection systems and refine the controls.” In exchange, “several of the world’s most influential figures will have access to early-stage technology designed to identify and manage AI-generated content that features their likeness, including their face, on YouTube at scale,” the streamer announced. CAA’s clients includes celebrity talent spanning acting, music and sports.

“CAA has a long history of advocating for artists’ rights, including significant investments in the CAAvault, a talent-focused service that scans, captures, and securely stores clients’ digital likenesses, including face, body and voice,” YouTube explained in a blog post.

In addition to managing AI-generated content that depicts faces, “YouTube has developed synthetic-singing identification technology within Content ID that will let partners automatically detect and manage AI-generated content on YouTube that simulates singing voices,” Variety notes.

In May, Variety reported that CAA recruited Veritone “to help build and operate” the CAAvault, describing it as an “early-stage database” housing digital replicas of its clients.

YouTube plans to add an update to the as yet unnamed takedown tech that lets “creators and rights holders choose to allow third-party companies to use their content to train AI models,” per Variety.

In addition to “highlighting AI-generated content using their likeness, this tool will give users easy access to removal requests through YouTube’s privacy complaint process,” notes The Wrap, adding that “the ultimate goal will be to make this technology widely available, thereby giving artists more awareness and control over how they’re being depicted on YouTube.”

YouTube is owned by Google, and it will be interesting to see whether the technology it is developing can eventually be applied to deepfakes that appear in Google Search.