YouTube has launched an AI avatar tool that lets creators generate photorealistic video clips of themselves for use in Shorts. The feature uses Google's Veo models and is rolling out globally, excluding Europe, for users 18 and older with an existing YouTube channel.
How the Avatar Tool Works
According to YouTube, creators record a live selfie and read voice prompts to capture their likeness and voice in a one-time setup process. After recording, they type text prompts to generate up to 8-second video clips starring their avatar. Multiple clips can be generated back-to-back for longer content.
The avatar can be retaken at any time, and creators can delete it permanently, which removes all recorded selfie and voice data from YouTube's systems. Avatars unused for three years are automatically deleted, though published videos remain unless manually removed.
The tool is available in both the main YouTube app and the YouTube Create app, with the same avatar working across both. No other user can generate content using someone else's avatar.
AI Transparency and Content Labels
YouTube has implemented multiple measures to indicate that content is AI-generated. All avatar-generated content carries mandatory AI disclosure labels. YouTube applies visible watermarks, Google's SynthID digital watermark, and C2PA metadata labels automatically during generation. These labels identify the content as AI-generated and cannot be removed by the creator.
YouTube says facial and voice data captured during the setup process is used exclusively for avatar creation. Deleting the avatar permanently removes the recorded data. The platform has not disclosed how long avatar data is retained in backup systems after deletion, only that front-facing deletion is immediate.
Building on Veo in Shorts
The avatar tool builds on YouTube's existing Veo-powered capabilities in Shorts. YouTube previously launched Reimagine, an AI remix tool that uses Veo to generate new visual variations of existing Shorts clips. The avatar tool extends the same underlying model into a different creation use case: generating original face-present content from text prompts rather than remixing existing video.
The practical effect is lowering the barrier for creators who want face-present content without being on camera. Avatar clips can be generated from a text prompt in seconds, bypassing the need for lighting, recording equipment, or physical presence. For brands and creators producing high-volume Shorts, the tool aims to reduce per-clip production cost.
Competitive Context
TikTok has integrated AI avatar capabilities through third-party tools such as APOB AI, but has no native first-party avatar generator built into the platform's creation flow. Snapchat launched AI Clips in March 2026, a Lens that transforms photos into five-second videos, but that feature converts static images into video rather than generating a personalised avatar from a live recording.
YouTube's native integration of avatar creation directly into the Shorts creation flow, powered by its own Veo model infrastructure, is structurally distinct from what other short-form platforms currently offer. The feature serves as both a creation tool and a creator acquisition mechanism, reducing one of the primary barriers to short-form video production: the requirement to appear on camera.
“This changes everything for business owners who know they should be creating more video content but hate being on camera constantly. The math is simple: your expertise × unlimited video creation = exponential reach. Instead of spending 3 hours a week recording, you invest 30 minutes creating scripts while your avatar handles the rest.
The businesses that figure this out first will dominate their space. While competitors debate whether AI content feels "authentic," early adopters will be capturing mind share at scale.”
Recap
How do YouTube AI avatars work?
Creators record a live selfie and read voice prompts to capture their likeness and voice. The setup is one-time. After that, creators type text prompts to generate up to 8-second video clips starring their avatar. Multiple clips can be generated and combined for longer content. The feature is powered by Google's Veo models and is available in the YouTube app and YouTube Create app.</p>
Who can use YouTube's AI avatar tool?
The tool is available to users aged 18 and older who have an existing YouTube channel. It is rolling out globally, excluding Europe. No other eligibility requirements have been specified beyond age and channel ownership.</p>
What AI transparency labels are applied to YouTube avatar content?
All avatar-generated videos carry mandatory visible watermarks, Google's SynthID digital watermark, and C2PA metadata labels. These labels identify content as AI-generated and cannot be removed by the creator. YouTube applies these automatically during the generation process.</p>
