Google Introduces Live Avatar with Gemini 3.8 Update
Google has unveiled a significant update to its AI capabilities with the introduction of the Live Avatar feature in the Gemini 3.8 Live model. This new update allows users to engage with an animated AI persona that can lip-sync and display various facial expressions in real time. However, this feature is currently exclusive to Gemini Enterprise customers.
Real-Time Multilingual Capabilities
The Live Avatar can seamlessly transition between 97 supported languages without compromising visual quality or causing video distortion. In a demonstration video, Google showcased the avatar conversing fluently in both English and Japanese, with precise mouth movements that synchronise perfectly with each language spoken. Additionally, the avatar has the ability to display information on-screen during interactions.
Customisation Options and Security Features
The news of Live Avatar’s launch comes shortly after Google announced the Gemini 3.8 Live model, praised for its near real-time processing of visual inputs. Customers have the option to select from a library of pre-designed avatars, or organisations can create bespoke avatars tailored to their specifications. Google ensures the integrity of these interactions by embedding an invisible SynthID watermark and implementing safeguards to protect identity.
Reporting based on coverage first published by theverge.com.