Google reportedly launched Gemini 3.8 Live’s Live Avatar feature
BIBIBI
AT A GLANCE
The evidence says Google launched a Gemini 3.8 Live feature with a real-time virtual avatar.
Article
Google goes live Gemini 3.8 Live’s Live Avatar real-time virtual avatar feature
Beating AI breaking news, Google released Gemini 3.8 Live last week and today launched Gemini 3.8 Live with Live Avatar: combining real-time voice conversations with low-latency streaming video generation, enabling agents in scenarios such as enterprise customer service and guided tours to “listen, see, and speak,” with lip synchronization, natural expressions, and smooth turn-taking. The feature is available on Gemini Enterprise as of today.
Live Avatar can process visuals and audio simultaneously, while conversations continue uninterrupted when tools are called asynchronously in the background, such as for hotel check-in. It supports seamless switching between 97 languages, with lip movements and expressions changing accordingly. Enterprises can use preset avatars or customize branded avatars using reference images (customization currently requires an enterprise whitelist). All audio and video outputs carry a SynthID watermark, making them identifiable as AI-generated content.
Original link https://m.theblockbeats.info/flash/368896
Key points
01
Beating AI breaking news reported that Google launched Gemini 3.8 Live with Live Avatar。
02
The feature combines real-time voice conversations with low-latency streaming video generation.
03
The report said the feature can be used in scenarios such as enterprise customer service and guided tours, and supports lip synchronization, natural expressions, and turn-taking.
04
The report said the feature is available on Gemini Enterprise as of today.
05
Live Avatar is described as being able to process visuals and audio simultaneously while keeping conversations uninterrupted when tools are called asynchronously in the background.
06
The report said the feature supports switching between 97 languages.
07
Enterprises can use preset avatars or customize branded avatars using reference images; customization currently requires an enterprise whitelist.
08
All audio and video outputs carry a SynthID watermark.
AI-assisted interpretation
The following is analysis, separate from reported facts. Verify important claims independently.
This is a feature that combines voice conversations with a real-time video avatar. Users can see a virtual avatar that speaks and can produce synchronized lip movements and expressions; enterprises can also use preset avatars or apply for customized branded avatars.
Why it matters to readers
If the information is accurate, agents in enterprise customer service and guided tours will be able to process voice and visuals simultaneously and respond to users in a way that more closely resembles face-to-face interaction.
Beginners can understand it as a real-time video agent capability that “can listen, see, and speak.”
Risks and unknowns
The information comes from Beating AI flash news; no official Google announcement or product page was provided as supporting evidence.
The scope of the enterprise whitelist, feature pricing, and available regions for custom avatars was not specified.
Enterprise customer adoption scale, actual latency, and 97 language support were not independently verified.
Related Developments
Loading event timeline…
Related concepts
real-time virtual avatar
This term is not in the glossary yet. Browse related concepts in the glossary.
