You've probably heard of Watson, the computer that went head-to-head with humans on Jeopardy. You know that robots are increasingly used in manufacturing around the country and around the world. But have you ever heard of a robot sketch comedian?
Well, meet Data. Also joining us is Data's handler, Heather Knight, a doctoral researcher in robotics at Carnegie-Mellon who studies the intersection of entertainment and robotics.
Check out the video for Data's comedy sketch, then read my interview with Heather Knight.
Fareed Zakaria: OK. So Heather, that was pretty amusing, but mostly just fascinating. Now, you wrote the routine for Data, but his reactions are sort of natural. His sensors work so that he can sense the audience's reaction.
Heather Knight: Right. Yes.
Fareed Zakaria: Explain how that works.
Heather Knight: So, like robots can learn through lots of data. In some of my work, I've been using each member of the audience as kind of a data point for machine learning. So in the reactions of a large group of people to a robot performer on stage, a robot could potentially learn to be more charismatic and more effective communicator and also be able to shape a performance for an individual group of people.
So there can be visual feedback, which is kind of conscious or we could make an iPhone app for your giving feedback along the way, like "I love that joke." You could rate things more, like Netflix style.
Fareed Zakaria: And – and the robot would, in effect, incorporate that information and tell more of the jokes that you like and fewer of the ones you don't – sort of like Pandora, with the thumbs up or thumbs down.
Heather Knight: Absolutely. Or you can even try telling jokes with a different set of gestures and see that joke is 10 times as funny for an audience.
Fareed Zakaria: Now, all of this sort of can be filed under artificial intelligence, and earlier this year, Watson, the IBM super computer, beat its human competitors in Jeopardy. So how sophisticated are we getting here?
Heather Knight: Well, I think that those two projects are actually great tandem projects. Watson is great at searching databases, and one of the things that I'm trying to do with the audience is generate some of those databases and also specifically generate them around social expression. So a machine can know how to actually communicate effectively with us, and so we don't have to adapt to using a screen or using a keyboard. They can learn how to work the way that we do.
Fareed Zakaria: Now, there are people, of course, who worry about something called the singularity. That is, the moment where robots will actually become smarter than humans, and will be able to learn and keep learning. Is that really going to happen?
Heather Knight: Do all parents feel that way about their children? I just wonder sometimes. I do feel like the way that we raise technology and the applications we use them for and the storytelling we think about in the creation of new technology will help us shape the direction that it's used. And we're not on the cusp of singularity at this very moment, but I do think that when you put people and robots together in teams, we can achieve much more than either of us can do alone. We're still very unique.
Fareed Zakaria: Heather Knight, Data, thank you very much.