Digital Duality: Google’s AI Gets a Human Touch in Gemini 3.8 Live

Google Introduces Live Avatar Feature in Gemini 3.8 Update

Google has rolled out a new feature in its latest Gemini 3.8 Live update that allows users to engage in conversations with a virtual model. This innovative “Live Avatar” feature enables users to interact with an animated AI persona in real time, complete with lip-syncing and various facial expressions. However, it is currently exclusive to Gemini Enterprise customers.

According to Google, the Live Avatar can seamlessly switch between the 97 languages it supports without compromising video quality or causing visual discrepancies. In a demonstration video shared by Google, the Live Avatar converses fluently in both English and Japanese, with its mouth movements accurately synchronized with the spoken words. Additionally, the Live Avatar can display on-screen information while engaging in dialogue.

The introduction of the Live Avatar feature follows closely after Google unveiled its Gemini 3.8 Live model, designed to process visual inputs in almost real-time. While Google offers a selection of pre-set avatars for users to choose from, organizations also have the option to create their own customized avatars. To ensure user privacy, Google embeds an invisible SynthID watermark in the output generated by the Live Avatar, along with protective measures to safeguard user identity.