Tech: You might be fooled while talking on a video call! This AI model has emerged..

What if an AI could step in for you during a job interview or a college viva and answer questions in real-time? San Francisco-based tech company Tavus has made this possible with its new Human Interaction Model, 'Griffin.' This AI tool is designed to facilitate face-to-face video conversations, complete with natural speech, facial expressions, hand gestures, and body movements. Unlike standard chatbots, this new AI tool not only listens to what people say during a conversation but also interprets their visual cues and responds instantly and appropriately. Let’s explore what Griffin is and how it works.

What is Griffin?
Griffin is an AI model developed by Tavus specifically for real-time video conversations. The company describes it as the world's first Human Interaction Model (HMI). Unlike typical chatbots that merely answer typed questions, Griffin is built for face-to-face interaction; it can hear your voice, observe visual cues, and respond using human-like speech, facial expressions, and body language.

Furthermore, according to the developers, the AI ​​model achieved a score of 3.83 out of 5 on NVIDIA’s VideoFDB benchmark for video generation. For comparison, real humans scored 3.92, while the next best system on the list scored only 2.80.

The company recently shared a video on X introducing this new Human Interaction Model (HMI), showcasing the AI ​​model in action. In the video, a person can be seen interacting with the model, which answers questions seamlessly without any interruptions or lag.

Notably, the responses are accompanied by natural facial expressions and hand gestures, making it difficult at first glance to distinguish whether you are speaking to an AI model or a real human. However, please note that we have not yet conducted an official test of this tool ourselves.

How does Griffin work?
According to the company, Griffin continuously processes the ongoing interaction, allowing it to react to interruptions, changes in tone, and other visual or auditory signals in real-time.

This model is built upon Tavus's existing research into video generation, speech, and conversational understanding. These systems enable Griffin to interpret facial expressions, speech timing, conversational pauses, and the surrounding environment.

The company believes that these capabilities make Griffin highly beneficial for applications such as mock interview practice, personal tutoring, and customer support; users do not need to learn specific prompts or commands but can instead converse naturally.

How can one use Griffin?
Currently, Tavus has not fully rolled out Griffin to the general public. The company states that it is working on disclosure and safety features before making the technology available to everyone. However, it is currently offering early access to "Griffin-Lite" to select testers and developers for initial testing.

Disclaimer: This content has been sourced and edited from Dainik Jagran. While we have made modifications for clarity and presentation, the original content belongs to its respective authors and website. We do not claim ownership of the content.