Google introduces Live Avatar for Gemini 3.8, adding animated facial responses
Google's Gemini 3.8 Live now features an animated avatar that lip-syncs and shows expressions while users chat, currently limited to Enterprise clients.
Google has rolled out a Live Avatar for its Gemini 3.8 Live AI model, allowing users to converse with a visual persona that mirrors speech through lip-sync and facial expressions. The avatar works across the model's 97 supported languages, maintaining video quality and avoiding visual drift, as demonstrated in a clip featuring English and Japanese dialogue. While the feature is presently offered only to Gemini Enterprise customers, Google will supply a library of ready-made avatars and permit companies to create bespoke versions.
The system embeds an invisible SynthID watermark in the output and includes safeguards designed to respect identity. Additionally, the avatar can display contextual information on screen while speaking, enhancing interactive experiences.
Why it matters
The feature adds a visual, multilingual interface to AI conversations, potentially reshaping how businesses interact with customers.
In this story
