Tavus unveils AI model that reacts like a human, holds real-time video conversations

San Francisco-based AI company Tavus has launched Griffin, which it calls the first "human interaction model" for real-time video conversations. The AI can see, hear, speak, interpret and react continuously, while adjusting its words, tone, timing and facial expressions based on the conversation. Griffin scored 3.83/5 on NVIDIA's VideoFDB benchmark, close to the human score of 3.92.