Google has introduced Gemini 3.8 Live with Live Avatar for its AI platform Gemini. With the new feature, AI conversations will not be limited to just voice, but users will also see a digital avatar, which will be able to listen, see and respond through voice and video in almost real time.
According to Google, Live Avatar has been designed in such a way that the conversation feels more natural. The digital avatar can show facial expressions while speaking and the movements of its lips will be in sync with the voice. The company says that this can make conversations with AI more interactive than before.
New feature is special for business
Gemini 3.8 Live with Live Avatar has been introduced mainly keeping business use in mind. Companies can use it for tasks like customer support, virtual assistant and interactive demo of product or service. This feature is currently available through Gemini Enterprise.
AI can see and hear simultaneously
The specialty of live avatar is that AI can process audio and visual input simultaneously. That means, answers can be given by understanding both what the user is saying and what the AI is seeing on the screen. While answering, AI uses facial expressions and video along with voice.
Google has also supported asynchronous tool calling in it. This means that AI can complete a task in the background while continuing to interact with the user. For example, it can keep talking to the user instead of stopping the conversation while checking some information.
Can communicate in 97 languages
Gemini Live Avatar is also designed for multi-language conversations. According to Google, this feature supports 97 languages. When the language changes during a conversation, AI can adjust its lip movements and facial expressions along with its voice according to that language.
Companies will also be able to create custom avatars
Businesses can choose from multiple preset avatars and different voices to suit their needs. Apart from this, organizations will also be able to create their own custom avatar. For this, by using high-quality reference images, an animated avatar can be created that maintains the look and style of the same character.
However, at present the ability to create custom avatars is limited to select enterprise organizations. Google has also used SynthID watermark to make identification of content created with AI easier.
