YouTube is expanding access to a tool that detects AI-generated videos with fake faces to journalists, government officials, and candidates for elected office, even if they aren’t part of the platform’s partner program.
YouTube is expanding access to a tool that detects AI-generated videos with fake faces to journalists, government officials, and candidates for elected office, even if they aren’t part of the platform’s partner program.
As 9to5Google reports, citing the YouTube blog, the company has begun a pilot expansion of its likeness detection system for public figures, who are often targeted by deepfakes. It is a tool that works according to the logic of Content ID, but instead of music or video fragments, it searches for facial matches of a specific person in generated content.
If the system finds an AI video with someone else’s appearance, the program participant can review such material and submit a request for removal. At the same time, YouTube does not remove all matches automatically. The company explains that each case is assessed individually for compliance with privacy rules. Exceptions remain for content that has public significance, as well as for satire and parody, even if it involves world leaders or other famous people.
To join the program, participants must go through identity verification through Google. YouTube emphasizes that the data users provide during this procedure is used only for verification and security purposes. According to the company, this data is not used to train Google’s generative AI models.
For now, access to likeness detection is open only to a limited group of journalists, government officials and candidates. YouTube says it plans to significantly expand the program in the future. At the same time, the platform called on American lawmakers to pass the NO FAKES Act, a bill that would enshrine a federal right to protect one’s public appearance in the era of generative AI.
In October, YouTube launched likeness detection exclusively for YouTube Partner Program members. The company is now testing the approach in the political and media spheres, where deepfakes can affect not only an individual’s reputation but also public trust in news and public communication.
Previously, dev.ua wrote about how YouTube is testing a new feature called Your Custom Feed, which allows users to independently create a home feed based on text queries, rather than relying solely on the algorithm, likes, and the «Not interested» button.
