Introducing Gemini 3.8 Live with Live Avatar
deepmind ·
Google DeepMind has introduced Gemini 3.8 Live with Live Avatar, an enhancement to its conversational AI that adds real-time visual presence through low-latency streaming video. This feature pairs near-real-time video generation with speech, allowing enterprise agents to listen, see, and speak using a dynamic visual persona with precise lip-syncing and natural expressions. The system supports asynchronous tool execution, enabling background data fetching while maintaining uninterrupted dialogue for complex tasks like hotel check-ins. It also features native multilingual support, seamlessly adapting lip-sync and expressions across 97 languages without degrading video fidelity. Organizations can utilize preset avatars or create custom ones that preserve brand identity and likeness, though custom creation is currently restricted to allowlisted enterprises. To ensure trust and transparency, all generated audio and video outputs are embedded with imperceptible SynthID watermarks to help minimize misinformation. The new capability is now available in Gemini Enterprise, offering richer and more accessible digital exchanges for customer service and interactive walkthroughs.
deepmind