YouTube has introduced a feature allowing users to create AI avatars that resemble them for use in Shorts. This functionality, which represents a continuation of Google Veo models, was previously previewed earlier this year.
The avatar creation process is accessible via the main YouTube app and YouTube Create. Users must capture a “live selfie” by recording their face and voice while reading a series of prompts. Each generated avatar will be photorealistic and may feature in YouTube Shorts for up to 8 seconds. Users can create multiple clips in succession, and only need to complete the setup once, although they can retake their selfie as needed.
YouTube stated that this new avatar feature allows users “an easier way to include themselves safely and securely in videos.” The video and voice data used for avatar creation will remain exclusive to the user’s avatar. Other users cannot employ the avatar for generating original Shorts. Users retain the option to delete their avatar at any time, with YouTube set to automatically remove avatars after three years of inactivity.
The platform noted that existing videos featuring a deleted avatar will still be available until those specific clips are removed. All avatar videos will include watermarks and digital labels, such as SynthID and C2PA, along with disclosures regarding AI-generated content.
To create an avatar, users should open the Create ‘+’ option in the YouTube app, tap the Gemini spark icon, and follow the prompts to select “Create video” and then “Make a video with my avatar.” The feature is also available in the Remix menu under Reimagine > Add me to this scene.
The rollout begins today for users aged 18 and older globally, with the exception of Europe. Full availability is expected in the coming days. Users must possess an existing YouTube channel to utilize this new avatar creation feature.
