en.Wedoany.com Reported - Google is expanding the Gemini Avatar feature to more paid Gemini subscribers, allowing users to create an AI digital version of themselves to generate videos and other content within Gemini. This feature is powered by Google's Gemini Omni model, the company's first video generation model that can combine multimodal inputs in a single prompt. First discovered in a March APK teardown, the feature aims to replicate the user's appearance and voice.

Users can access the feature by navigating to "Settings > Avatar" in the Gemini app. Once enabled, users can create a personalized AI avatar that mimics their facial features, voice, and mannerisms, and summon it using specific commands in Gemini conversations. The setup process is relatively straightforward: the first phase requires a camera-based training session where users hold their phone at eye level and follow on-screen prompts to capture their face from multiple angles, allowing Gemini to generate a detailed model of facial structure. The second phase is voice training, where Gemini asks users to read aloud a series of random phrases and numbers to analyze their speech patterns, tone, accent, and rhythm. After setup is complete, the avatar is linked to the user's Google account and becomes available within the Gemini experience.

To ensure facial and voice scans are accepted, Google recommends users keep their phone at eye level, avoid overly dark or bright environments, and ensure their eyes, nose, and mouth are visible. The system allows glasses as part of the avatar but prohibits hats, sunglasses, and masks, and no other people or facial images should appear in the user's background.

Addressing potential risks of realistic AI avatars, Google has announced several safety measures. Users must be at least 18 years old to create an avatar, and the account owner must be physically present during the registration process. Additionally, all videos generated using Gemini Avatar are embedded with Google's SynthID watermark, an invisible watermark directly embedded in the generated content that can be detected via Google tools, helping viewers identify AI-generated videos and reduce the risk of misuse. Currently, Google has not announced whether this feature will also be available to free Gemini users.

This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com