Gemini 3.8 Live Adds Lip-Synced AI Avatars With 97-Language Support

Google has launched Gemini 3.8 Live with Live Avatar in Gemini Enterprise, adding near real-time video personas to its live conversational AI models. The feature combines speech with generated video so enterprise agents can respond with synchronized facial expressions, lip movements, and spoken dialogue.

Live Avatar builds on Gemini 3.8 Live, which Google introduced last week. The new version gives the model a visual presence that can react while listening and speaking, creating a more interactive interface for uses such as customer service and guided walkthroughs.

The system processes audio and visual input at the same time. That allows an avatar to respond to what it hears and sees while maintaining the flow of a conversation.

Google also designed Live Avatar to keep talking while other tasks happen in the background. Through asynchronous tool execution, the system can call external tools, retrieve information, and complete backend actions without stopping the active conversation.

One example shown by Google involved a hotel check-in interaction, where the avatar continued speaking with a guest while handling supporting tasks in the background.

Live Avatar also supports multilingual conversations across 97 languages. Google says the system can adjust lip-sync and facial expressions as a conversation shifts between languages without degrading the video output.

Enterprises can choose from preset avatars or create custom ones. Developers can generate a personalized avatar from a high-quality reference image while preserving the appearance, branding, or identity represented in that source. Google says custom avatar creation is currently limited to organizations approved through an enterprise allowlist.

The company is also adding transparency measures to the generated output. Audio and video created by Live Avatar include SynthID, Google’s imperceptible watermark for identifying AI-generated content.

Google says the watermark is embedded directly into both the audio and video, with the goal of making generated material easier to detect and reducing the risk of misattribution.

The release moves Gemini’s live enterprise agents beyond voice-only interactions. Instead of operating as an unseen assistant, businesses can now deploy a responsive visual character that speaks, reacts, switches languages, and continues working with connected tools during a conversation.

Gemini 3.8 Live with Live Avatar is available now through Gemini Enterprise.

This analysis is based on reporting from Google.

Image courtesy of Google.

This article was generated with AI assistance and reviewed for accuracy and quality.