YouTube is expanding its AI-powered deepfake detection tool to help politicians and journalists track fake videos of themselves on the platform.
The feature, called likeness detection, helps identify videos that use artificial intelligence to recreate a person’s face or appearance without permission.
The tool has already been available to millions of YouTube creators. Now the company is testing it with a small group of journalists, government officials, and political candidates to see how it works for people who are often targets of deepfake content.
Likeness detection works similarly to YouTube’s Content ID system, which scans the platform for copyrighted music or videos. Instead of looking for copyrighted material, this system scans videos for a person’s face. When the system finds a possible match, the person enrolled in the program receives a notification and can decide whether to request the removal of that video.
However, not every removal request will automatically be approved. YouTube says each request will be reviewed under its privacy rules. The platform still wants to protect freedom of expression, which means content such as parody, satire, or political commentary may remain on the platform even if it includes AI-generated versions of public figures.
Leslie Miller, YouTube’s vice president of government affairs and public policy, explained that protecting open discussion is an important part of the platform. According to her, parody and political criticism are often part of public conversation, and removing them automatically could limit free speech.
To join the program, individuals must submit a video of themselves along with a government identification document. YouTube says the information will be used only for the likeness-detection system and that participants can leave the program at any time and request that the company delete their data.
Amjad Hanif, vice president of creator products at YouTube, said that even though the system detects many matches, the number of removal requests from creators has been relatively small. In many cases, the detected content turns out to be harmless or even helpful for the creator’s online presence.
For politicians, however, the situation may be different. AI-generated videos can spread misinformation quickly, especially during election periods. Because of this risk, the platform is exploring different ways to manage deepfake content responsibly.
Interestingly, YouTube is also considering the possibility of allowing monetization on some AI-generated deepfakes in the future. Hanif suggested that some creators and industries may eventually want to allow this type of content under controlled conditions.
The rise of artificial intelligence has created new challenges for online platforms. Over the past few years, companies like YouTube have struggled to manage AI-generated music, voices, and videos that imitate real people.
Deepfake technology has become a concern not only for celebrities and public figures but also for everyday individuals. However, YouTube says expanding the likeness detection system to everyone is not currently part of its plan. For now, the tool will mainly focus on well-known figures and people frequently appearing in news coverage.
That said, regular users can still report AI-generated deepfakes through YouTube’s existing complaint system if they believe a video violates their privacy.
At the same time, artificial intelligence is becoming deeply integrated into the YouTube ecosystem. The company has introduced many AI tools designed to help creators brainstorm ideas, plan videos, generate content, and optimize their channels.
AI-generated videos are already appearing across the platform. Some reports have shown that children are sometimes exposed to low-quality AI videos that claim to be educational but offer misleading or inaccurate information.
If this article helped you, please consider supporting our work. Every small contribution keeps Abijita.com independent and running.
To address these concerns, YouTube has recently taken action against several channels producing low-quality AI content. Some channels with millions of subscribers have been penalized under spam policies, and others creating fake AI movie trailers have also faced restrictions.
