Google launches Gemini Avatar for creating a digital twin of the user

Google has begun a wider rollout of the new Avatar feature in the Gemini app for paid subscribers.

The tool allows users to create a digital AI copy of themselves based on their appearance and voice. The feature is powered by Google's new Omni model.

Every created video contains an invisible SynthID watermark, and the feature is available only to users aged 18 and older.

During the Gemini Avatar setup process, there comes a moment when everything stops feeling like a typical AI capabilities demo and starts looking quite unusual.

The user looks into their smartphone camera while Gemini instructs them to slowly turn their head from side to side and speak random numbers. After a few seconds, Google creates a digital version of the person using their face and voice. Users can then see videos in which the avatar speaks things the user never recorded themselves.

Google began expanding access to the Avatar feature for paid Gemini users several months after the capability was first discovered in an APK analysis of the app in March.

You can create your own Gemini Avatar in the Gemini app via the "Settings" → "Avatar" section. The feature uses the new Omni model to recreate the user's appearance and voice, after which it allows generating content featuring them.

After completing the setup, the avatar can be invoked directly in Gemini chats using commands like @me or the user's name.

Users must record their face and voice by following step-by-step instructions. The system will ask them to look into the camera, turn their head in different directions, and speak certain numbers aloud to accurately reproduce facial features and voice.

Source: androidauthority