YouTube Unveils ‘Likeness Detection’ to Combat Deepfake Videos
In an effort to curb the rise of artificial intelligence–generated fake videos or “deepfakes,” YouTube has introduced a new feature called Likeness Detection. The technology will identify when a user’s face has been altered or used without permission through AI and ensure the removal of such videos from the platform, according to a report by Engadget.
In its initial phase, the feature is being rolled out for members of the YouTube Partner Program. At present, the system can only recognize facial likenesses; it cannot yet detect when a person’s voice has been manipulated by AI.
To use the feature, interested users will be required to submit a government-issued ID and a short video selfie to verify their identity and store their facial data. Once registered, the system will function similarly to YouTube’s Content ID technology—scanning videos across the platform to identify those that use a person’s face without consent. Users will then have the option to report and request removal of such videos.
Analysts describe this as a significant step by YouTube in tackling the growing misuse of AI-driven technologies.
