Key Highlights:
- YouTube Shorts now supports AI avatars that look and sound like creators.
- Setup requires a one-time live selfie video and voice prompts.
- Avatar clips can generate videos up to 8 seconds per prompt.
- All AI avatar videos include SynthID, C2PA labels, and disclosures.
YouTube is rolling out a new AI avatar feature for YouTube Shorts that lets creators generate videos using a photorealistic version of themselves. The avatar copies both face and voice after a quick setup inside the app. The feature is launching globally outside Europe for users aged 18 and above.
This marks another step in bringing generative AI deeper into YouTube Shorts creation workflows.
What Is the New YouTube Shorts AI Avatar Feature?
The new YouTube Shorts avatar tool allows creators to appear in videos without recording traditional footage. Instead, users generate short clips using prompts after completing a one-time capture process.
Creators record a live selfie and read voice prompts. Then the system builds a digital version of them. After setup, the avatar can appear in multiple Shorts without repeating the process.
Each prompt generates clips up to eight seconds long. However, creators can chain multiple clips together to build longer videos.
How Does the Setup Process Work?
The feature is available inside the main YouTube app and YouTube Create.
Users can access it through:
- the Create “+” button
- the Gemini spark icon
- Remix > Reimagine > Add me to this scene
Once inside, creators select “Make a video with my avatar” and enter prompts describing the scene. The avatar then appears automatically inside the generated clip.
Users can retake their avatar capture anytime if their appearance changes.
How Is Google Expanding AI Inside YouTube Shorts?
The avatar tool builds on earlier video generation features powered by Google Veo models already used in Shorts. Previously, creators could generate videos using uploaded images. However, voice-matched avatars introduce a new level of personalization.
According to YouTube, the goal is to help users “include themselves safely and securely in videos” without complex production steps.
This makes AI video creation faster and more accessible.
What Safety Protections Are Included?
YouTube says selfie recordings and voice samples are used only to create avatars. Other users cannot reuse them.
Creators can delete avatars anytime. Meanwhile, unused avatars are automatically removed after three years. However, videos already published will remain until manually deleted.
Importantly, every avatar clip includes transparency labels. These use SynthID watermarking and C2PA metadata standards to identify AI-generated content clearly.
Where Is the Feature Available Now?
The rollout starts globally outside Europe. It applies only to users aged 18 or older who already own a YouTube channel.
As AI tools continue expanding inside the platform, this update shows how YouTube Shorts is moving toward prompt-driven video creation with built-in identity protection.
