Certain videos created for parody or satire may still be allowed, YouTube says.
YouTube is expanding its AI likeness detection tool to allow political candidates, government officials, and journalists to request the removal of their deepfake videos.
Launched last year in YouTube Studio, the tool lets users find videos in which their faces appear to have been altered or generated using AI. To sign up for the program, you need to be a YouTube channel owner or manager and submit a government-issued ID and a short selfie video for verification. YouTube will then search for videos featuring your likeness and list them under the Content detection tab in YouTube Studio. You can review each video and request the removal of any you find manipulative, or request the removal of all of them.
YouTube, however, warns that it won’t take down all videos. The company says it will continue to protect free expression and allow parody or satirical content, even if they criticize world leaders and influential figures. “We’ll continue to carefully evaluate these exceptions when we receive requests for removal,” it adds.
For now, YouTube will start with a “pilot group” of unnamed journalists, government officials, and candidates who are getting early access to the tool.
AI deepfakes remained a major point of discussion throughout the 2024 US presidential election. There were reports that foreign governments were trying to manipulate voters by spreading deepfake videos of former VP Kamala Harris on social media. We also saw fake celebrity endorsements for then-candidate Donald Trump.
AI detection won’t be enough to tackle the spread of deepfakes, YouTube admits, adding that it continues to support the NO FAKES Act. The bill would require online platforms to remove unauthorized digital replicas of a person upon request from the rights holder.
