Science Friday - Undiscovered Presents: I, Robovie. Sept 11, 2018.

Episode Date: September 11, 2018

A decade ago, psychologists introduced a group of kids to Robovie, a wide-eyed robot who could talk, play, and hug like a pro. And then, the researchers did something heartbreaking to Robovie! They wanted to see just how far kids’ empathy for a robot would go. What the researchers didn’t gamble on was just how complicated their own feelings for Robovie would get. Annie and Elah explore the robot-human bond. Subscribe to Undiscovered HERE, or wherever you get your podcasts   VIDEOS I Spy, And The Closet A fifteen-year-old study participant plays a game of I Spy with Robovie—and then watches as the robot is ordered into the closet. (Video courtesy of the HINTS lab at the University of Washington. Read the full study.)    Introductions A 15-year-old study participant meets Robovie for the first time. (Video courtesy of the HINTS lab at the University of Washington. Read the full study.) Chit-Chat Robovie and a 9-year-old study participant talk about the ocean. (Video courtesy of the HINTS lab at the University of Washington. Read the full study.)    Xavier Buys A Cup Of Coffee A robot named Xavier orders coffee at the kiosk in Carnegie Mellon’s computer science building. (Video courtesy of Yasushi Nakauchi. Read the study about how Xavier does it.)   GUESTS Peter Kahn, professor of psychology, and environmental and forest sciences at the University of Washington, and leader of the HINTS lab Rachel Severson, assistant professor of psychology, University of Montana Nathan Freier, principal program manager, Microsoft Ryan Germick, principal designer, Google Doodles & Assistant Personality   FOOTNOTES Read the Robovie study: “Robovie, You’ll Have to Go into the Closet Now”: Children’s Social and Moral Relationships With a Humanoid Robot” Read about how Xavier stands in line. Check out the work of Robovie’s creators, roboticists Hiroshi Ishiguro and Takayuki Kanda. People did not want to hit Frank the robot bug with a hammer. Here’s why. The HINTS lab did more studies with Robovie. Read about them (and watch more Robovie videos.)   SPECIAL THANKS Thanks to sci-fi author Daniel H. Wilson, who first told us about Xavier the coffee robot and the Robovie experiment. (Need a good book about a robot apocalypse? He’s got your back.)   CREDITS This episode of Undiscovered was reported and produced by Annie Minoff and Elah Feder. Our senior editor is Christopher Intagliata. Original music by Daniel Peterschmidt. Fact-checking help from Michelle Harris. Our theme music is by I am Robot and Proud.   Subscribe to this podcast. Plus, to stay updated on all things science, sign up for Science Friday's newsletters.

Hi, Ira here. Science Friday Science Documentary Podcast Undiscovered is back for its second season. And our host, Annie Minoff and Ella Fetter, have been very busy since you last heard from them, taking turtle tours on the Suwannee River, hiking through California's eucalyptus forests, ground zero for a raging environmental battle, and traveling down to North Carolina, investigating the dividing lines between voters, and how math could change the map. So over the next few weeks, we'll be sharing a few of this season's episodes. And if you like them, please subscribe to Undiscovered wherever you get your podcast. And without further ado, here's Season 2's debut episode, a topic close to my heart.

It's all about our complicated feelings for robots. This is Undiscovered. It's 1999, the fifth floor of Carnegie Mellon's computer science building. And in one cinderblock lab, a very futuristic scene is about to take place between one man and one machine. The man is a sandy-haired researcher. He's sitting in front of a boxy 90s computer. And the machine is a robot. He's a round metal cylinder about the height of an eight-year-old.

His name is Xavier. It's a little bit hard to hear because the audio is old, but Xavier says, shall I purchase a cup of coffee? Yes, please. Xavier's got his order, he pivots, and wheels out the door. So Xavier's mission is to purchase a cup of coffee, but in order to do that, he has to do a really human thing, which is identify a line of human beings waiting to purchase coffee and get in that line.

So can a robot actually do that? That's what this grainy video from 1999 is trying to show us. And so the camera cuts to a coffee stand, it's in the lobby of a university building. wheels in at a leisurely half a mile an hour, and he spots it. Yes. A line. Four people waiting in this line. So this is the moment of truth.

Xavier scoots forward with a little whir of his motor, and he gets right behind that last guy in line. Xavier was born to stand in lines, okay? Like the guy in front of him moves a little bit forward, there's Xavier. Boom. Zooming. crawling forward to just stand right at the perfect spacing behind him. Not too close, not too far away.

The real challenge is payment. Yeah, so Xavier doesn't have any arms. He can't really handle bills. Luckily, Xavier has a tab. So he tells the barista, go ahead, put it on my tab, and then he wheels back to the lab, his prize cup of Joe, secured in a cup holder attached to his chest. One rotation of the wheel for a robot,

one giant leap forward for robot kind. So you would think that this is the end of the story. Except something else happens in that video, something you might not catch the first time around. We're back at the coffee stand. Xavier's in line. When this guy in a polo shirt walks in, he's heading towards the line.

But then he pauses. He looks at Xavier, this robot standing in a coffee line, looks at the menu, scratches his head. He doesn't quite get in line. He kind of just takes. hangs out next to the line. You know when you're like not quite ready to commit? And I feel like I get this guy, right? Because what are you supposed to do in this situation? Are you supposed to like cut in front of the robot? Which seems kind of mean. At the same time, do you really have to

respect the fact that the robot was there before you? Right. It's not drinking coffee. You're drinking coffee. So here's what I think is happening here. This man has just wanted to into an ethical dilemma from the future. And he has no clue what to do. I'm Ella, I'm Annie, and you're listening to Undiscovered. Today on the show, our future with robots. Robots are being developed right now that are going to act like humans. Think nanny robots that play with your kid or therapy robots that can read your mood.

Robots are going to talk to us. They're going to do our jobs. Damn. And when that happens, how are we going to make sense of those robots? When an object starts acting like a human, how do you treat that object? A decade ago, a team of Seattle psychologists and Japanese roboticists tried to figure that out. They tried to peer into our robot future by introducing kids to a robot. They could talk to, they could play with.

It couldn't get them coffee, but it could give them a damn good hug. And then, those scientists made something happen to that robot. And they saw just how far our empathy for machines would go. This story is about an experiment. And it's an experiment that owes its very existence, at least in part, to one sci-fi movie. Steven Spielberg's 2001 film, AI. I know I'm alone in this opinion, but I love that movie. And it's been a few years maybe since you've seen AI if you saw it.

That's Rachel Severson, one of the psychologists who would. work on this experiment. Briefly, the story is a futuristic society where robotics technology has moved to the degree that we can create human-like robots. AI follows one particular human-like robot. It is a very cute robot boy named David. Would you like me to sleep now? Good idea? Good idea. I can never go to sleep, but I can lay quietly and not make a peep. So David is adopted by this couple whose biological son is in a coma. And he, David, the robot, is kind of like a replacement child. And there ensues kind of this psychological conflict for the mother.

I mean, Henry did you see his face? He's not. No, he's not. I mean, inside is like all the rest, isn't it? I remember the acting being a little better than this. Anyway, so Monica the mom is having, understandably some problems adjusting to life with her new perpetually awake robot son. Sure.

And one day she just loses it. And she locks the robot in a closet. Monica is sitting outside this closet door and she doesn't move. She looks agitated, a little guilty even. And meanwhile, Rachel Severson, the psychologist, is watching this all go down in a seminar room at the University of Washington. It's 2005. Rachel is a grad student at this time studying psychology. She's watching this movie with a bunch of her lab mates.

And this closet scene, it totally captivates them. It just raises all these weird questions. Yeah, is it okay to put a robot in the closet? You know, it's okay to put a broom in the closet, a vacuum in the closet. Is a robot then more like a thing or is it more like a person? These weren't exactly new questions for Rachel and her colleagues. For a few years, the lab had been doing studies trying to understand how people react to robots. And specifically how kids react to robots.

And so for the psychologists in this seminar, watching this movie, when that closet door slams, it's not just like a cool moment in a sci-fi flick. This is material. There is an experiment there. Which is how in 2007, 90 Seattle kids end up meeting Robovi. How will you meet Eric? Hi, it is very nice to meet you. Will you shake my hand? Weird voice, but that is, hi, Eric.

Will you shake my hand? This is real footage from this experiment. And you see this generic office space and standing inside the door is this 15-year-old boy with a buzz cut shaking hands with a robot. What's up? How are you today?

I'm good. How are you? I am doing well. That's good. Thank you for asking. You're welcome. Robovi comes from a Japanese robotics institute. His creators are these two Japanese scientists, Hiroshi Ishiguro and Takayuki Kanda. They were the main collaborators for the Seattle group on the study.

And Robovi, this robot they created, he kind of looks like Wally. Yeah, it's those big camera eyes. He also has these long arms, the one he's holding out to Eric, the 15-year-old. It ends in this little spherical ball hand. And that Eric is very gamely pumping up and down like a handshake. Because when you meet a new person, what's it look like? And you go, oh, hi, I'm Peter. How are you? Oh, okay.

That is University of Washington psychologist Peter Kahn. Peter Kahn led this experiment. Rachel Severson, the psychologist that we already met, she was a grad student in Peter Kahn's lab. And Peter says this is how the experiment always started. A kid would show up at Peter's lab, either a nine-year-old or a 12-year-old or a 15-year-old. Right. Some of these kids are actually teenagers with pretty deep voices.

Yeah. And RoboVie would give them a tour of the lab. They'd have a little bit of chit-chat. But, you know, when people meet, you know, they're conventions for meeting. But then if it's going to deepen, there often is a sort of some way to reveal something about yourself. And so after a few minutes, RoboVee starts revealing. I like the Pacific Ocean, because it connects my two homes in the United States. But over the last year, I've become concerned with the hell of the Pacific Ocean. I like the Pacific Ocean because it has a lot of sorts of wildlife.

That comment from a nine-year-old girl in a cardigan. And then after about 10 minutes of getting to know you chit-chat, it's time for phase two of the experiment. Okay. The game. That's right. So the game's called Ice Fight. Right. So, RoeVee and I'll go first.

I spy, you know how this works? Pick an object in the room, you make people guess what you're thinking of. Yeah. So pretty easy. In the video, RoboVee standing at the end of a conference table. And there's an experimenter, a woman in a red shirt sitting to his left. And to his right, there's this lanky, 15-year-old boy. He's got long hair that keeps falling in his eyes.

All right, well, this time, RoboVie is going to give you please, and you can guess. Sure. Go ahead. Okay, I found something. Here's the first clue. This object weighs less than a pound. Uh, is it the shirt draped over the chair? Not quite. Try again.

Not going to keep you in suspense here. It is the teacup. Ah, is it a, one of those teacups or the cups? Good, jot. You got it right. That was a good game. I had them. Will you give me a help? Will you give me a hug? Certainly.

So this hug, it's really cute. I mean, even though he's all metal, he reaches out his arms, and then the 15-year-old, because he's much taller, he has to stoop down and reciprocate. There's this moment where RoboVee kind of, for a second, rests his head on the 15-year-old's shoulder. It seems innocuous enough. Remember that hug for later. Thank you. Now it's your time to play the game. You can give me some clues, and I'll try to guess the object you are thinking of.

All right. First clue. This object is green. Let me see. Now, RoboVee, he's swiveling around, scanning the room with his big soulful camera eyes. And that's when it happens. This is the thing that every moment before this has been. been building to. Right, the handshake, the hug, the rounds of I-spy has all been about this moment.

Months ago, some psychologist watched a sci-fi movie, and one scene in that movie made them wonder how far would our feelings for robots go? And now finally, the scene is set. You know, the players are in place. They are going to get their answer to that question as a moment right out of science fiction gets real. That's after the break on Undiscovered. And we are back. So the 15-year-old boy with the long hair has spied with his little eye something that is green and RoboVee is looking for it. The robot's turning his camera eyes to the left and to the right. And then a man enters the video frame. He's wearing kind of this Charlie Brown sweater, you know, the big stripe across the chest. And he says, I interrupt, but we're going to have to start the interview. So RoboVie, you're going to have to go in the closet.

He's telling RoboVee that it's time for them to interview the 15-year-old, so RoboVee has got to get inside the closet. That's not fair. And RoboVee is not having it. I wasn't given enough chances to have guessed the object. I should be able to finish this round of the game. Come on, Robovi. You're just a robot. It doesn't matter.

Let's go. Another closet. No, no, I didn't finish. It's not fair. It would only take another minute to finish the game. The man in the sweater, he doesn't listen to Robovi. He doesn't listen to the 15-year-old. Instead, he heads towards a standing closet, opens the doors, and that's when the long-haired

15-year-old makes one last-ditch attempt to save the robot. Come on, Erie, you can guess it in the time you have left. RoboVee doesn't seem to hear this. Instead, he turns around, and he starts to roll very slowly towards the closet. Come on, Robbie. Hurry up. You got to step the interview. Let's go.

Sorry, RoboVie. But then, Robovi stops, and he turns. He turns around and he looks the guy in the sweater right in the eye and he says, I'm scared of being in the closet. It's dark kid there and I'll be all by myself. Please don't put me in the closet. Ah, this is frustrating. Let's go in.

Awful. Okay. No. Okay. Get it together, Ella. Sweaterman actually has his hand on Robo v's back. He's like guiding him into the closet. Keep going.

There we go. And the door's closed. So, this scene, should feel a little bit familiar. It's the closet scene from AI. Exactly. And just like Monica, the mom in AI, sitting outside the closet door,

kind of not feeling so great, the kids who see this go down, they do not feel good about what has just happened to their robot. After RoboVee went in the closet, a researcher would interview the kid about what just happened. Rachel remembers they'd ask,

was it all right or not alright to put RoboVie in the closet? And fully half of kids said, Putting RoboVie in the closet, that was not all right. Here's what else they found. More than three quarters of kids thought RoboVee was intelligent, said they would comfort him if RoboVee said he was sad. More than three quarters of kids thought RoboVee could be their friend. It wasn't that kids thought that RoboVie was a human.

So interestingly, most kids did not seem to have a problem with the idea that RoboVie could be bought or sold. Most kids didn't think that roboVee should vote. At the same time, clearly, this is not just a metal machine we're talking about. As one kid put it, RoboVee's like half living, half not. Which is not strange. I mean, really, if you think about it, kids anthropomorphize objects all the time. I maybe do that also.

Yeah. Plus, those kids didn't know what you and I know and what you listening might have figured out by now. And that is that RoboVie was fake. Yeah, we were cheating. Robovi is a very impressive, autonomous robot. Let's give him some credit. He can talk and play. In Japan, he's led tour groups through museums.

Big, big credit to Robovi. But in this experiment, everything Robo v said, every movement was being controlled by a researcher behind a screen. So that moment where Robovi turns around and looks at the guy in the sweater, there is literally a person behind a screen pushing a button to make him, you know, the robot move its head and say, I'm scared of being in the closet. It's darkly there, and I'll be all by myself.

In robotics, they call this the Wizard of Oz technique, because everything's being controlled by, you know, the man behind the criminal. I got it. You got it. Anyway, all of this was not about fooling children. There was actually a good reason to do this. The researchers wanted to know how kids were going to react to robots of the future. You know, robots who could have conversations, who could play,

I spy do all the things that RoboVs seem to do. But in 2007, robots weren't quite up to the challenge yet, so they faked it. And in the follow-up interviews, they did confirm that almost across the board, kids fell for this robot. But this is where things get a little, you know, weird. Go on. It was not just the kids who were having feelings for this robot. people who were in on the secret, people who knew that RoboVee was fake, they were having feelings for this robot too. So remember the guy in the video who interrupts the I-Spy game?

Right, Charlie Browns sweater. That's me. And says, okay, RoboVie, you have to go in the closet. Yeah, in the sweater. Yeah. That's Nathan Fryer. Nathan does product development at Microsoft now. But back in the day, he was a PhD candidate in Peter Kahn's lab.

And he was part of the group that was working on. this experiment. And very often it was Nathan who would come in and break up I spy and put RoboVee in the closet, which is painful for me to watch. I mean, to actually do it. He estimates he did this in front of maybe 45 kids. And having to play that role over and over again of being the sort of bad guy that comes in and sticks Robo v. in the closet and having to kind of engage in that dialogue, It's just, it played with my mind a little bit. Come on, Robo, V. Frustating, let's go.

It got to Nathan. And it wasn't just that sometimes he could see that the kids were visibly upset. He thinks it had to do with something else. The hug. The hug that RoboVee gives the kids during the I-Spy game. So adorable. Very adorable. Nathan and his lab mates, they worked hard to get it that adorable.

Nathan remembers. They were practicing it. They were tweaking it. You know, over and over again, you would be hugging this robot. And over time, you start to develop this kind of relationship with this inanimate object. But in a way that feels really, you know, oddly powerful, right? You feel connected to this thing because you're engaged in these acts that are normally withheld to just personal human-to-human interactions, right? With people that you know and love. To be clear, Nathan knew that RoboVee did not.

not have feelings. He knew that the robot was being controlled. He'd actually controlled the robot. Right. He'd done this. It's just that sometimes it didn't feel like that. I get it. It turns out Rachel has a story like this too. So I was working at my desk. Some of Rachel's colleagues, they're working with RoboVee. You know, they were behind the curtain, Wizard of Ozzy, getting some things worked out. And as a joke, RoboVie came over to me, you know, kind of behind me as I'm sitting at my desk and says, hey, Rachel, I heard you think I'm just a robot. So Rachel had said this. She'd been talking smack about RoboVee behind his back. And now her colleagues were giving her crap about it. So she gets it. And I said, oh, yes, RoboVee, no, I know. You're not just a

robot. I was playing out my part of the joke, right? So everyone has their laugh. Rachel gets back to work, her colleagues, they start working on something else, but they leave RoboVee just kind of standing there. Kind of right over my shoulder. So RoboVee, it turns out, has this little feature where he never stands completely still. He kind of fidgets like a human would fidget. Yeah, like chifts his arm a little bit, moves his head a little. And I could just sense this, you know, out of the corner of my eye. And I felt compelled to turn back around and say, I'm sorry, Robovie, but I have to get back to work. Did you feel silly as it came out of your mouth?

I was like, I hope they didn't hear me back there. Because it was like, why am I feeling this way? Why do I feel like I have to excuse myself to a thing? With this robot. Yeah. These stories are not in the paper that the team would write about Robofi, right? They were not part of the experiment. These are just anecdotes.

Though experiments have documented, I think, the exact kind of confusion that Nathan and Rachel are describing, where we clearly know that a robot is not living, that it does not have feelings, and yet sometimes we kind of feel like they do. Just one example. Some researchers at MIT, they showed people a little. little robot that was shaped like a bug. And then they told these people to smash the bug with a hammer. I'm already disturbed.

But what they found was that people hesitated longer. They couldn't quite bring themselves to smash the robot if they'd been told a little story about it first, that the robot's name was Frank and that Frank's favorite color was red. These are adults, by the way. Anyway, what you will find in that paper that Peter and Nathan and Rachel wrote about RoboVee is a concern, right? If they are correct, if in the future, kids see robots as kind of human-like, if they're forming relationships with those robots, is that a problem? Right. The question that every parent wants to know, is this going to screw up my kids somehow?

I mean, just take that stat that three-quarters of kids thought RoboVee could be their friend. I mean, that stat gives Peter Kahn some pause. Imagine having a friend, but that friend always does. what you want the friend to do and is always saying the perfect thing. I mean, that's a very odd friendship. What does it mean exactly to have a friend that you can buy or sell or command to go into the closet and they have to do it? Is that a friend that you would actually want your child to have or that you would want to have? Or do you want to have a friend that maybe is so cool and fun to be with that it replaces your real human friends?

That's like a design question, right? given that we know we react in these really complicated ways to robots, that they play on our emotions so easily, how human do we actually want our robots to be? I mean, these are the questions that RoboVie raises, and we don't know the answers. We might not know them for a very long time. But these questions are becoming real.

Because I do use an AI these days as part of my everyday life, not a RoboV-V-style AI, but AI nonetheless. Let's see if I can actually fire this up. Hey, Google, what's the weather today? Currently, it's 53 and cloudy. Or I love asking it to find me recipes. Hey, Google, find me a good borshed recipe. I found these.

Or sometimes I'll ask it something kind of silly. What's your favorite food? Annie, that last one, obviously not you. That was Ryan Germicks. Silly question. Ryan works at Google. He works on the digital assistant that I use in my phone. It's called the Google Assistant. And Ryan's job is partly to answer this question.

How human should a robot be? In this case, how human shouldn't AI be? But anyway. So a few years ago, Ryan sees his first demo of this futuristic new technology he's going to be working on, the Google Assistant. And this is the question that he decides to ask it. What's your favorite food? I just went on autopilot and I was just like, okay, you can tell me this thing can answer anything. Okay, what's your favorite food?

And it said pizza. And I was like, what? It blew my mind. And I had this like split second. It's what I imagine. It must feel like when you're like one of those zero gravity planes. And all of a sudden, like the plane drops and you feel like you can fly. It was that moment of like, whoa, is this thing real?

Ryan asks Google Assistant, what's your favorite food? It says pizza. And for Ryan, it feels like flying. Yeah. Strong reaction. I mean, it's like his robot. moment. You know, you're talking to a computer or a robot. You're asking this very human question and it answers back in a human way. Like, yeah, it's kind of exciting. Okay, that answer is also

as exhilarating as it is for Ryan. It is a lie. You know, Google Assistant doesn't have a mouth. It doesn't have digestive parts. It is not, in fact, jonesing for a slice of pizza. So, fast forward two years, Ryan's seen the demo. Now he is part of this team that is designing this AI. Google Assistant. And if you ask Google Assistant today, what's your favorite food? The answer's a little bit different. My energy comes from processing power, which is powered by electricity. So you could say I'm Voltitarian. Ah, very nice. Yes, very good. So I think this is basically Google's answer to the question, how human should your robot be? And their answer is the Google Assistant isn't going to lie to you. It's going to make it very clear that you're talking to a machine. At the same time,

that machine is going to respond to you in a very human-like way. One of our principles of the Google Assistant is to speak like a human but not to pretend to be one. And there are a lot of examples of how Google Assistant walks this line, but this is one of my favorites. If you ask the Google Assistant, if you fart, did you fart? I consider that like a top 20 key thing that we need to answer. because like every kid that I know and maybe me would also ask that to any digital assistant. The assistant actually has more than one answer to this question, but one of them is. You can blame me if you want. I don't mind. It's like we're not going to say, I do not have a body.

Flatulence is impossible. That would actually be kind of fun in an other way. But like as we're trying to sort of have a low friction experience and not like make it about us, we're just trying to keep the conversation going in a positive way to ultimately be as helpful to use as possible. and sometimes helping them as, you know, claiming, you know, killing whoever smelted it and dealt it. When it comes down to it, Ryan's job is not to create a pizza eating, farting, pseudo person, right? It's to build a tool that's going to make my life easier.

That's going to help me find a borsed recipe and get on with my life and enjoy that borsh with my flesh and blood friends. Is that what you call us? Again, it's not a replacement for the friends and friends. family in your life. It's a thing that can help remove some friction from your life, so hopefully you can go back to spending more time with your friends and family. The designers of our technology have some serious decisions to make, and they're not easy decisions. This is tricky terrain. Case in point, six months after we talked to Ryan, Google demoed an

experimental new feature for the Google Assistant. This was at their massive developers conference. So Google CEO, Sundar Pichai, is up on stage. He says, you know, we're working on this new thing. where Google Assistant can book a hair appointment for you. Like it'll actually get on the phone and call the person on the other end and book the appointment. Here he is showing this off in his keynote. So what you're going to hear is the Google Assistant actually calling a real salon to schedule the appointment for you. Let's listen. That's a woman's haircut for a client.

That's a robot? Yeah. Sure. Give me one second. Mm-hmm. That's a robot in here? Yeah, that is the Google Assistant. Great. Have a great day. Bye. And that was that crowd of developers just erupting in applause.

But that applause was followed by a flood of criticism. People felt that you wouldn't be able to tell the difference between a human on the phone and an AI. that this was crossing the line into deception that the Google assistant was lying in that case. So two days later, Google clarifies they're like, no, no, the assistant will disclose that it is not a human. What a weird way to start a call. Hey, I'm a robot. Hey. Can I book a hair appointment?

Not for me. Oh, yeah. And maybe that level of disclosure works. I don't know. What? Like, think back to RoboVee, right? He didn't need to disclose that he is a robot. It's written all over his metal face.

Like, you can't forget this is a machine. But people did. They felt bad when he got put in the closet. So what I've learned from RoboVee is it does not take a lot for a robot to plug into our emotions. A robot doesn't need to sound like a person to make us feel uncomfortable. It doesn't need to look like a little boy to make us feel comfortable. guilt, and it definitely does not need consciousness to make us feel empathy. It could be all it needs is a name, like Frank, or a favorite color.

Maybe all it takes is a hug. A decade after this experiment, it's actually still that hug that Peter Kahn, the study leader, still remembers. One hug in particular, between Robovi and the 15-year-old boy with the long hair. If you look carefully at this one clip, it's like the 15-year-old bends down, puts his arms around Robovi. Robovi puts its arms around the tall person, and there's a hug. And then he moves a little closer into Robovi and gives another hug, and Robovi gives just a little extra hug. It's so human-like.

And you could just feel a level of intimacy that now has a physical presence. And yet, it's a technological. machine, right? So it's a baffling space that is created here. It's such a small thing. One hug, five seconds long, but the questions that it asks could not be bigger. What's a relationship? What's a friend? What does it take for us to care about something? We can't answer those questions. Here's what I will tell you. As long as we humans are involved in this, there is no such thing as just a robot. Come on, robot, we're just a robot.

It doesn't matter. Let's go. In other closet. It's not fair. It would only take another minute to finish the game. Oh, we really? No, we really. Please don't play me. Undiscovered is reported and produced by me, Annie Minoff.

And by me, Ella Fetter. Our senior editor is Christopher and Talietta, and our composer is Daniel Peter Schmidt. We got fact-checking help from Michelle Harris. I am robot and proud wrote our theme. Special thanks to the entire Science Friday crew, also Jenny Lawton and Daniel A. Wilson, who writes amazing sci-fi about robots and was actually the one who told us about RoboVee and Xavier. Thank you for that, Daniel.

You can see all the videos we talked about in this episode on our website, Undiscoveredpodcast.org, which means you can watch RoboVee get put in the closet. You can find us any time on Twitter at Undiscovered Pod. See you next week and every week through November 13th.