Google adopts AI avatars on YouTube Shorts as OpenAI discontinues its Sora video generation platform.
YouTube Shorts is launching a new AI-powered feature that enables creators to clone themselves on camera. This addition, previously hinted at, highlights the platform’s complex relationship with AI-generated content. The new tool allows users to create an avatar for insertion into Shorts videos or to generate new ones. These avatars, which mimic the user’s appearance and voice, are designed as a secure method to use AI for content creation.
To create an avatar, users must complete a “live selfie” that captures their face and voice with specific prompts. YouTube recommends optimal conditions such as good lighting and a quiet environment for best results.
Users can use these avatars to create videos up to eight seconds long by selecting “make a video with my avatar,” as noted by 9to5google. Avatars can only be used in the creator’s original videos, and creators can manage the remixing of their Shorts. Avatars not used for three years will be deleted.
Avatar videos will have visible AI-generation indicators like SynthID and C2PA. The feature is rolling out gradually, requiring creators to be at least 18 and have an existing YouTube channel.
This joins YouTube’s suite of AI tools, such as AI-generated videos on Shorts, AI auto-dubbing, and AI analytics, many powered by Google’s Gemini AI models. These tools allow users to create videos, music, and images.
The introduction of avatars comes as OpenAI withdraws from video generation. OpenAI recently announced discontinuing its Sora video tool, citing financial and legal challenges.