San Francisco artificial intelligence startup Tavus has launched Griffin. The company describes Griffin as the first Human Interaction Model. This class of model aims to understand and hold face-to-face conversations in real time.
Griffin processes multiple communication signals simultaneously. It handles speech, facial expressions, tone of voice, gestures, and pauses. The system manages these tasks while both receiving and generating video.
Study results show high realism
In a company study, 48 percent of participants believed Griffin was a real person after a one-minute video call. Previous systems achieved a maximum of only two percent.
Griffin scored 3.83 points in an independent Nvidia test measuring how human an AI feels during direct audio-video conversation. Actual humans scored 3.92 points in the same evaluation. The previous best artificial intelligence model reached 2.80 points.
Company background and rollout
Founded in 2020, Tavus has raised about $64 million in funding. The startup initially created personalized artificial intelligence videos for sales and marketing before expanding into live digital persona conversations.
A preview version named Griffin-Lite is currently available to select testers as a research preview. A more capable version will launch once safety concerns are addressed. Tavus suggests potential use cases include tutoring, practicing difficult conversations, and camera-based tech support.
More details on Griffin are available in the Tavus research report.
