YouTube has officially switched on a new set of AI-powered Shorts tools that let creators generate short videos featuring a photorealistic avatar of themselves. After previewing the feature earlier, the platform is now making it available so eligible users can produce AI-generated Shorts that look and sound like the creator, without having to film every clip in the traditional way.
To get started, creators complete a one-time setup designed to capture their likeness and voice. The process includes recording a “live selfie,” where the user films their face and reads a series of prompts. YouTube uses that data to build an AI avatar that can appear on camera in Shorts. Each AI-generated clip can be up to eight seconds long, and creators can stitch multiple clips together to make a longer Short. The avatar doesn’t need to be rebuilt each time, but YouTube says it can be updated later if a creator wants to refresh how it looks or sounds.
YouTube is also putting guardrails around who can use these avatars. Creators can only generate videos using their own avatar, and they can’t use someone else’s likeness to produce Shorts. On the data side, YouTube says these avatars are automatically deleted from its servers after three years of disuse, adding a defined retention window for accounts that stop using the feature.
Once the avatar is set up, creators can generate videos with simple text prompts, powered by Google’s Veo text-to-video technology. The idea is to make it easier to brainstorm, prototype, and publish Shorts with less time spent filming, while still keeping the creator’s on-screen presence intact.
The rollout begins globally today, April 09, with Europe excluded for now. The feature is available in the main YouTube app and YouTube Create, and it’s limited to users aged 18 and older who already have YouTube channels.
YouTube has framed the update as another step in making AI a creative assistant rather than a replacement for real creators. The company says the goal is to help people experiment with new formats, including AI-generated video and music, while keeping the creator’s identity at the center of the content.
At the same time, YouTube knows this kind of tool sits in the middle of a larger debate about deepfakes and synthetic media. AI-generated videos of real people can be convincing enough to fool viewers, and the growing presence of realistic AI content has made it harder to tell what’s authentic. That can lead to hoaxes spreading faster—and it can also cause real videos to be dismissed as fake.
To address those concerns, YouTube says it will clearly label content created using its AI tools so viewers understand when a video is AI-generated. The company also emphasizes that creators must disclose when they’ve made realistic altered or synthetic content. In addition, YouTube has introduced an AI deepfake detection tool aimed at helping politicians and journalists, positioning it as part of a broader effort to reinforce trust on the platform.
Even with labeling and detection tools, the bigger challenge remains: as AI video creation becomes easier and more accessible, separating reality from fabrication will continue to be an ongoing issue for creators, platforms, and viewers alike.
