Meet Griffin, the first AI model that can interrupt you just like a real human being
Is it real or is it AI?
Image by Tavus
Follow us
- Tavus says Griffin is a human interaction model designed for realistic video conversations.
- The company says 48% of participants believed Griffin was a real person.
- Demos show Griffin laughing, showing surprise, interrupting, and reacting to being interrupted.
Key Takeaways by nexos.ai, reviewed by Cybernews staff.
Research lab Tavus has released what it calls a human interaction model, or HIM, claiming that it’s “the first AI to pass the video Turing test.”
The company said 48% of participants who talked with Griffin during the Turing test believed they were talking with a real human being.
Currently, Griffin-Lite, which the company calls a research preview, is available to a select group of testers, with its “more powerful” model planned to be released for wider use in the future.
What can Griffin do?
The company presented the model by showing how it talks with real human beings, as if they’re on a video call with another person.
Multiple demos show how Griffin is capable not only of answering questions, but it can also show certain expressions, behaviors, and emotions as it responds to what’s being said by laughing or showing surprise.
The model also boasts a realistic image that doesn’t get distorted even when it's being asked to wave, touch its face or tuck hair behind an ear, making the whole interaction a lot more real.
Another example video shows how realistically the model reacts to a conversation by interrupting or being interrupted.
Griffin was also shown to help with certain tasks, such as solving a Rubic’s cube, work as a practice partner, or help with simple tasks, such as following time and directions during a project.
Tavus shared that such a model could impact how “people learn, work, or get help” by being turned into a tutor, project buddy, or always-available listener.
“I think we should pass a law making this illegal”
The company shared its findings on its X account, prompting users to discuss how useful the new type of assistant may be.
Despite how realistic-looking Griffin is, many were concerned about its use.
“Now we don't have to show up to remote job interviews anymore. It's only going to get weirder and weirder from here on,” wrote one X user.
“I think we should pass a law making this illegal. I’m ok with real-time cartoon characters though,” added another netizen.
Some users shared that they can still see that it’s not a real human behind the screen.
“Forty-eight percent is basically a coin flip. Past this point, the only reliable way to know you're talking to an AI is if it tells you, so disclosure stops being optional,” an X user said.
