Have you ever had a conversation with a computer that wasn't one-sided? Virtual humans are the newest attempt to cross the uncanny valley and present individuals with computer-generated, interactive characters that can empathize, understand, and exhibit emotions. Artificial intelligence has come a long way, and research at the University of Southern California's Institute for Creative Technologies is pushing the limits of what's possible in modern computer science.
I sat down with William Swartout, the Institute's Director of Technology, to discuss the future of virtual human interaction. He also let me play in their virtual saloon, where I got into a gunfight that was just a little too real.
Click the link below and/or watch the video above to learn more. And don't forget to leave a comment! Talk nerdy to me!
CARA SANTA MARIA: So what's a virtual human?
WILLIAM SWARTOUT: So a virtual human is a computer-generated character that is designed to look and act as much as possible like real people. And the idea is that these characters are autonomous. That means that they're thinking on their own. They're using AI [artificial intelligence] to figure out what's going on in their environment and respond appropriately.
UTAH: Howdy.
CSM: What's your name, bartender?
UTAH: My name's Utah.
CSM: Are you from Utah?
UTAH: I moved here from Gatlin, Tennessee many, many years ago.
CSM: So these characters aren't just programed to go through a script.
WS: No.
CSM: They're actually responding to my actions in the environment.
WS: Exactly. Exactly. And so if you do something different, they're going to respond differently.
CSM: So how do they know where I'm standing?
WS: So they're using computer vision, and there are actually infrared cameras that are in the ceiling of this set that are capturing your position and tracking where you are. There's a marker on the hat you were wearing that, sort of, points out what direction you're pointing. And that can be used to understand which character you're talking to at any particular point. So if you say, you know, "what's your name?" and you're pointing to one character, that's the character who'll answer.
CSM: And also I noticed some of that reflective tape on my gun.
WS: On your gun there, yup.
CSM: So is that how they know when I've drawn the gun?
WS: They know when you've drawn the gun, and they know where the gun is pointing.
CSM: Oh yeah.
WS: So if you're pointing at, you know, at the wrong character and you pull the trigger, bad things are going to happen.
CSM: [laughs] So this is kind of a fun game, I guess. What are the real world applications of this?
WS: One of the applications for the Army--who are actually our major funder--is to create scenarios and environments where people can rehearse missions that they might be going on. So it's doing that kind of training that allows people to possibly make mistakes in simulation before they actually have to do things in reality.
HARMONY: Oh no. He's here!
CSM: Rio's kinda hot.
HARMONY: Shh. Be quiet!
CSM: So what do you think the future holds for virtual humans?
WS: Well, you know, I think in its broadest terms, this is really going to be a new way of interacting with computers. I think that a new metaphor, maybe the next metaphor for interacting with computers is going to be something a lot like virtual humans. We're already starting to see that with things like Siri. And what virtual humans add to what we've already got in terms of human-computer interaction is really a familiarity. Interacting with a virtual human is much like interacting with another person. So you don't have to learn a lot of stuff--
CSM: It's intuitive, like the Wii! But more intuitive than the Wii.
WS: It's very intuitive. Exactly. Exactly so. And, the other thing that virtual humans bring is a social element to the interaction. So when you're interacting, there's a lot of studies that show when you're interacting with a virtual human, people respond to virtual humans in much the same way that they respond to real people. And so you can actually--and we've demonstrated this--you can actually build up rapport with people.
CSM: So this is, kind of, one of the big vistas in computer science right now, is it not?
WS: Absolutely. I mean, I think this is really a whole new, very interesting area that's opening up. And the whole idea of creating virtual humans where they're really thinking for themselves is a huge intellectual challenge.
RIO: Where's that lowly wagon wheel? Utah?
CSM: Utah?
RIO: Shut the hell up, right now! Utah? You spineless coward, get out here. Where is he?
HARMONY: He didn't say.
RIO: Why the hell are you still here?
CSM: I'm here to defend Harmony's honor.
RIO: I hope you said your prayers today, Ranger. 'Cause you're about to die.
UTAH: Rio--
RIO: Utah, what the hell are you doin'?
HARMONY: Now! Shoot him, Ranger!
[gunfire]
CSM: [laughs]
CSM: Well, thank you so much for having this chat with me and inviting me to your saloon!
WS: Thank you. Thank you for being here.
See all Talk Nerdy to Me posts.
Like Cara Santa Maria on Facebook.
Follow Cara Santa Maria on Twitter.