YouTube Shorts is giving creators a way to put themselves in videos without filming. The platform began rolling out an AI avatar tool on April 9 that builds a digital clone of a creator based on their face and voice, then places it in new Shorts or inserts it into existing ones.
The rollout comes as YouTube has been expanding AI creation tools for creators while also contending with an ongoing problem with AI slop, deepfake scams, and impersonation content across the platform.
How creators build and use an avatar
Building an avatar starts with a recording called a “live selfie.” Creators follow a series of on-screen prompts while the system captures their appearance and voice. YouTube recommends filming in good lighting, in a quiet space, with no other people or facial images visible in the background, and with the phone held at eye level. The company says the resulting avatar will “look and sound like you.”
Once built, the avatar can generate new Shorts of up to eight seconds from a text prompt, or be inserted into existing videos already in the creator’s feed. YouTube has not specified what criteria a Short must meet to be eligible for avatar insertion, describing the requirement only as “eligible Shorts.”
Restrictions and content labeling
YouTube has placed firm limits on what creators can do with their avatar. The tool only works within the creator’s own original videos. Creators control whether those videos can be remixed by others, and they can delete their avatar or any video featuring it at any time. Any avatar left unused for three years will be automatically deleted.
Every video that includes an avatar must carry AI-generated content labels. YouTube says these include visible watermarks, SynthID digital tagging from Google, and C2PA metadata. C2PA is a broadly supported content authentication standard that embeds information about how content was created, though its effectiveness depends on platforms and devices preserving the metadata after content is shared or re-uploaded.
Rollout and eligibility
The feature is rolling out gradually. YouTube has not given a timeline or specified which regions will get access first. Creators must be at least 18 years old and have an active YouTube channel to participate.
YouTube’s broader AI toolkit
The avatar tool adds to a growing set of AI features Google has built into YouTube for creators, including AI-generated clip suggestions for Shorts, automatic video dubbing in multiple languages, and a channel analytics chatbot. Most of these tools run on Google’s Gemini AI models, which also support photo-to-video generation, music creation, and image synthesis.
The rollout arrives about a month after OpenAI wound down its Sora video generation platform. OpenAI cited high costs alongside a persistent stream of copyright disputes, deepfake controversies, and low-quality generated content. The shutdown came as OpenAI was preparing for an anticipated IPO.