NEAL CONAN, host:
This is TALK OF THE NATION. I'm Neal Conan, in Washington.
In Woody Allen's 1973 film "Sleeper," Allen's character wakes up from a deep freeze 200 years in the future. The world is run by a Big Brother-type authoritarian government where robots do much of the domestic work. Allen disguises himself as a servant robot with a silver painted face and a dome on his head and gets fitted for a suit by two robotic, yet very Jewish tailors. It's a satiric take on the play that gave robots their name: Karel Capek's "R.U.R.," which stands for Rossum's Universal Robots. You might be more familiar with "The Terminator."
These dystopias all portray thinking, walking and talking machines that are a very long way from the automatons that actually exist today. Think of the robots that manufacture cars on assembly lines. But in turn, those robots are a long way from the cutting edge, where some robots are barely beginning to act independently. Our guest this hour reports that he was startled on a couple of different occasions when he unexpectedly ran into robots that were at least in some respects almost human. Writer Lee Gutkind spent six years off and on as a fly on the wall at one of the country's leading robotic labs, watching the work of the gear heads and code monkeys who work tirelessly to build robots that will move, see, learn and even understand.
Later in the hour, after Hitler, Stalin and Mao, amid Darfur and Iraq, Harvard Professor Steven Pinker argues that we are probably living at the most peaceful time in history.
CONAN: But first, robots, technology, and the future. If you have questions about the state of the art, where it's going, and what it means, our number is 800-989-8255, 800-989-TALK. E-mail: email@example.com. There's also the conversation on our blog: npr.org/blogofthenation.
Lee Gutkind joins us from the studios of member station WFAE in Charlotte, North Carolina. His new book is "Almost Human: Making Robots Think." Good to have you on the program.
Mr. LEE GUTKIND (Author, "Almost Human: Making Robots Think"): It's nice to be here.
CONAN: And you bumped into Grace one night while you were outside a hotel, and for a few seconds, you wondered what was going on.
Mr. GUTKIND: It scared me to death. I did - I was not outside a hotel. I was in the corridors in the middle of the night at the Robotics Institute...
Mr. GUTKIND: ...at Carnegie Melon, and suddenly a creature confronted me and asked me why I was roaming the corridors. What are you doing here, the creature said. Now I had spent a number of years interacting with robots, but at that - and I knew as I looked at that robot, I knew that it was a robot, but, you know, it scared me to death. It surprised me the way in which it confronted me, just for that second, and I was really flummoxed. I was startled.
CONAN: Interesting. You write that, for example, how we humans interface with computers, this has been a respected branch of science for some years now. How we interface with robots, that's just beginning.
Mr. GUTKIND: This is just beginning now, you're right, because robots, of course, are just beginning. We've been speaking about robots, and if you watch the science fiction movies, the ones that, say, Steven Spielberg produces, we think that robots are just like us, but in fact robots are far away from where we are today. And it is just now beginning to become important to think about robots and think about how human beings are going to interact with robots, because in 25 or 30 years, we're going to be forced to do just that. Even now, in hospitals there are nurse-bots which deliver medications and roam the corridors of hospitals doing small but important tasks. And in Japan there are robots that are help — beginning to help human beings, elderly people, take care of themselves. And there are robots now, at Carnegie Melon in fact, that act as receptionists, greeting people as they walk into the door.
CONAN: Yet one of the things that characterizes the search for a robot - and this is something you spend an awful lot of time on - is how do you decide when a robot is autonomous, when it's making independent decisions about the right thing to do?
Mr. GUTKIND: Well, there's a great range of autonomy, and sometimes roboticists get incredibly excited and feel incredibly positive and say that their robots are being autonomous, which of course what roboticists want. They want to be able to create and - a figure to create a creature that is very much in their own shadows.
Mr. GUTKIND: But robotists - there are levels of autonomy. So sometimes perhaps a robot can walk across a room, go through a door, and do it autonomously. But then there might be a creature standing right in front of that robot. The robot might be able to go around that creature, but the robot will not know what that creature is, will not be able to recognize it or perceive it or know if it's a man, a slice of pizza, or a rope dangling from the ceiling. And so that's not necessarily where we are with the way humans are autonomous, but it's a beginning.
CONAN: Mm-hmm, one of the things in fact you write about, and most interestingly, is as you look at what is necessary to make an artificial creature do some of the things that we do naturally, you begin to be astonished at just what wondrous beings we truly are.
Mr. GUTKIND: That's exactly what I was just saying in relation to everyday things that we do. Here I am. I'm sitting and talking into this microphone, chatting with you. My hands are going in all kinds of different directions. I'm taking notes. I'm using the fingers in my hands. There are bones and tissues and nerves. And robots are not - are pretty much unable to do most of that. Robots can speak with you sometimes, but we are such incredible creatures. I can walk down to the bank and get some cash out of a cash machine. Robots can't do that yet. These are just normal, everyday things that we do, and the best that a robot can do in many respects is vacuum a floor or scrub a window. And so we are very much far away from the multiple, multitasking human being perspective.
CONAN: Of course, eventually we'll progress to the point where robots, of course, won't do windows, but that's another story. If you'd like to join our conversation with Lee Gutkind, give us a call: 800-989-8255, 800-989-TALK. E-mail us: firstname.lastname@example.org. Let's begin with Ed. Ed's calling us from Kansas City, Kansas.
ED (Caller): Yeah, hi, everybody.
ED: I was calling in more with a comment than a question. I work as a middle school mentor with the Kansas Robot League here in the state of Kansas, and we utilize the Lego MindStorms robots as adjunct teaching tools to bring children up in the - what are referred to as the stem disciplines: science, technology, engineering, and mathematics. And we are emphasizing working particularly in inner city schools where we have girls and minority representation, as a means of encouraging them to follow along in those careers, college degrees and careers. The Kansas Robot League was started by my younger brother as he was working on his doctorate in education, and now we've spread it into a number of schools in basically the eastern half of the state of Kansas.
CONAN: Interesting you should mention that. Lee Gutkind writes about how robots will inspire students much more than teachers or even books, but also he - the way you're writing, Lee Gutkind, those students that Ed's working with in Kansas City, Kansas, they're the - they're going to be the shock troops of the robotics industry in, well, just a couple of year time.
Mr. GUTKIND: This is what is so - was so amazing to me at Carnegie Melon, that so many - I envision roboticists as someone who looks like me, that is to say -somewhat elderly, wild white hair, kind of maybe looking like an Einstein kind of figure. But most of the people doing the cutting edge work in robotics today are maybe not middle school people, but are people under 30, and it reminds me of the movement in Mountain - in Silicon Valley at the end of the 1960s and early 1970s when people started to - when young people started to investigate that thing called a computer. And it's the ideas that were stimulated and the experiments that were done by young people that brought us to the modern computer age of today.
Mr. GUTKIND: And the same thing is happening with robotics now. And so I think it's wonderful that - but not necessarily unusual - that middle school students are turned on by robots. And this is also - robotics is becoming an international language. It's not just happening in the United States. It's happening in South America, it's happening in Australia. Robotics is kind of moving all of our young cultures together in one, major, strong, forceful movement that's going to make robotics technology explode over the next quarter of a century.
CONAN: Hmm, you write extensively about several different projects in your book - mostly about the people rather than the robots, I do have to say. There's one in particular I wanted to ask you, and that was a robot called Groundhog which some kids - under the direction of one of your older, wild-haired leaders, a charismatic guy - they ferociously put this robot together to go into a mine -this was after the Quecreek disaster in Pennsylvania...
Mr. GUTKIND: Yes.
CONAN: ...to go into and map a mine. This was something that a robot could do, they thought, where a person wouldn't be safe. And it's a story of enormous frustration and all the things that can go wrong, but ultimately it's a story about success, too.
Mr. GUTKIND: That's true, and that's kind of what robotics seems to be presenting today. The challenge is so great - and we know a little, we don't know a lot - and so there - here was a group of undergraduate and graduate students - most of whom, as I said, were under 25, led by this charismatic guy, the father of field robotics, a man named Red Whitaker. And they had this idea that if the Quecreek Mine had been properly mapped, the water would not have flooded the mine and people would not have been so endangered. And so they decided to create a robot and write the software and try to learn - get a robot to autonomously - now remember, we can't go into the mine, a flooded mine, with the robot - to get the robot to autonomously map the mine. And the creativity in just building the robot was really incredible. They started out by going to a golf cart junkyard...
Mr. GUTKIND: ...and they found a frame for a golf cart that looked like it was strong enough and small enough to hold the computers and the hardware necessary to go into a mine. And they built it from the ground up. This is just not eggheads who write software. This is kind of a combination of gear heads, engineers, and hardware people and software people. And they built this robot, and the robot creeped - when I say creeped, these robots think but they - kind of think - but they think really slowly. So they creep for a few yards, they stop, they creep for a few more yards, they look to the left, they look to the right, and then they go forward. And that robot, that one robot, the first time made it about halfway through this experimental mine, this possible flooded mine, before it stopped dead. And I'll leave it right there. It stopped dead, and the students were trying to discover whether that robot was going to come out or not.
CONAN: Stay tuned for the exciting adventures of Groundhog in our next episode, coming up right after the break. It's TALK OF THE NATION from NPR News.
(Soundbite of music)
CONAN: This is TALK OF THE NATION. I'm Neal Conan, in Washington.
We're talking today with Lee Gutkind about his new book "Almost Human: Making Robots Think." You can talk with Lee yourself. Give us a call: 800-989-8255. E-mail us: email@example.com. And just before the break we left the robot Groundhog stranded in an abandoned mine, halfway through. Lee Gutkind, what happened?
Mr. GUTKIND: The scene, just a little bit more. These nine kids are sitting in a boxed panel truck, with their computers on their laps, and they're watching the heartbeat of the robot as it goes deeper and deeper into the mind. And suddenly the heartbeat disappears. And the smiles and the glory that was on their faces just disappeared instantly. It was one of the saddest moments I can remember in my life. And the robot was missing for an hour, and they lost it for an entire hour. And they thought - they really did think that that robot, that they had worked so hard on, would disappear forever. But a mine official, official from the Bureau of Mines, went into the robot - went into the mine and found the robot, and guess what happened, Neal?
Mr. GUTKIND: The computer crashed.
(Soundbite of laughter)
CONAN: No kidding.
Mr. GUTKIND: All he had to do - that was it.
Mr. GUTKIND: All he had to do was press a button, reboot it. Groundhog came back out, and so they kind of thought this was a somewhat of a success. But in subsequent tries, the robot, Groundhog, went in and out, in and out, and successfully and continually mapped the mines. And this kind of presents what the future, the potential future, of where - what we could do with robots. You know, so many mine accidents are caused because mines are not properly laid out. We don't quite know where the water is and what's happening between the walls, so...
CONAN: Mm-hmm, yet as you also point out at the end of the book, there have been a lot of mine problems since that, and nobody has used a robot.
Mr. GUTKIND: Because the robots are like Groundhog.
(Soundbite of laughter)
Mr. GUTKIND: They're just not ready for prime time, and people are not putting the money, the resources into such research in order to polish it and make it happen. And frankly, Neal, we're talking about maybe a three to five-year scope of possibilities. Just because we pour millions of dollars into robotics doesn't necessarily mean that we're going to move out of these shadowed ages and into the age of robotics. It's not going to happen for a little while. But it's never going to happen unless we begin to think and to understand that the future of progress in this world will very much rest on the world of robotics.
CONAN: Let's get some callers on the line. We'll turn to Mike. Mike's with us from Oakland, California.
MIKE (Caller): Yes, thank you for taking my call.
MIKE: My question has to do with the - what's happening with communication between robots and humans? Just as a quick example, companies have IVR systems, and I've had very bad luck with them understanding the responses I give when I'm forced to talk. I mean I literally yelled an epithet at one when I became frustrated. And it said, oh, you want customer service. So I'd be worried about, you know, when your guest said that he was confronted by a robot in the hallway that said what are you doing, I'd be afraid about how the robot would understand the answer and what it would do about it.
Mr. GUTKIND: I was afraid, too. And in fact we're way behind. If we're behind - we're moving ahead with technological research as far as getting robots to become more autonomous, but this human computer, or human robot interaction field, is so far behind. And, you know, I mean that's kind of what happened with the whole world of computers and the Internet. We are not - we're allowing technology to explode in front of our eyes without planning for what's - for the impact of technology. And how the technology is going to interact with human, we are not giving a lot - enough attention to that very important aspect.
MIKE: It sounds like - it doesn't sound like the front is too good right now.
Mr. GUTKIND: The front?
MIKE: It doesn't sound like progress is very good right now in that respect.
Mr. GUTKIND: It is just this human robot interaction, this field is brand-new. It is just beginning. People are just now beginning to understand how important it is for humans and robots to get along. And of course, as you well know, pretty soon, very soon, as robots progress, robots and humans are going to be working side to side in factories - I mean in space the whole scenario begins with a robot going into space and clearing the way for humans. And when humans arrive on Mars or on the Moon, they are going to have to work very, very closely with - kind of as supervisors and quality control officers for the robots on the scene. And if we don't know how to communicate with our robots, we're going to have a lot of difficulty.
CONAN: Mike, thanks for the call.
MIKE: Thank you.
CONAN: Here's an e-mail from Eric, in Cincinnati. Can your guest speak to the law that was recently passed in Britain to protect the rights of robots, should they become more ubiquitous in our lives?
Mr. GUTKIND: No.
(Soundbite of laughter)
Mr. GUTKIND: I think that's - I did not - I am unaware of that law and...
CONAN: I was going to say it's a surprise to me, too.
Mr. GUTKIND: Yes. So I guess I don't know what to say except that it's unclear - it really is unclear - the difference between robots, the potential of robots, and human beings is in many ways unclear. I mean the subtitle of my book is "Making Robots Think," and it's very unclear whether even human beings actually think. Robots are following a set of rules that human beings are writing for them, and in fact kind of thinking that most human beings do on any given day in many respects follow a different set, but also a set of rules as well.
CONAN: Mm-hmm, let's turn to Dave, Dave with us from Sterling Heights in Michigan.
DAVE (Caller): Hi, I was wondering if you'd looked at the possibility of - or projections of when robots or computers that operate on the robots would approach human intelligence, based on Moore's laws, Moore's Law of...
DAVE: ...of microprocessor doubling every year and a half.
CONAN: Yeah, and I know that there are people at MIT who think that they're going to overtake human intelligence within this century. But, Lee Gutkind, the way you write about robots in the year 2050, one of the goals of one of your teams was to build a team of robot soccer players who could beat the World Cup champions by 2050.
Mr. GUTKIND: I'm telling you, that was - actually, it was RoboCup, this whole RoboCup movement that got me involved in writing this book initially. When I learned that groups of robots were playing soccer against one another, teams of robots, in this RoboCup league - which was launched in I believe 1996 or 1997 -I looked into why. And the whole goal of the - this RoboCup movement is to be able to develop teams of robots that can be skillful enough physically and also clear enough emotionally and intellectually to beat the pants off of the World Cup champions, human beings, 50 - by the year 2050. And to me that was really quite a remarkable idea, and I laughed until I saw the robots playing soccer in action.
So you have to kind of look at this, Neal, in two different ways. These robots cannot play soccer as good as human beings, good human beings can play soccer. But on the other hand, right now robots can play soccer, they can move a ball back and forth, move a soccer ball - or a smaller ball - back and forth across a court. They can pass to one another, they can work as a team, they will know when they're winning, and when they're winning they will go into a defensive mode. And when they're not winning, they'll become quite aggressive. They can be put into the penalty box and 30 seconds later be put back into the penalty - into play and know - have a great awareness of what the score is and what their teammates are thinking. So this is quite amazing.
CONAN: You used a word in that description that a lot of people are going to have a problem when you're talking about robots...
Mr. GUTKIND: Yes.
CONAN: ...and that's emotions.
Mr. GUTKIND: Yes. Well, I think that that's - robots, of course, we think are creatures without a soul, without spirit, and I think that that is really what the distinguishing factor is between human beings and robots. If in fact robots will not develop this spirit, this soul, this competitive edge, then probably human beings will be able to maintain the lead both in robotics and in life - I mean in soccer and in life for a long time to come. But if indeed somehow robots are able to grasp this inside, this soul that human beings have, I think we have a different story entirely, and...
CONAN: Yeah, one of your - one of the people you quote in the book said, look, if it's a matter of speed and power and flexibility and athleticism, we think eventually the robots are going to beat the World Cup champions. If it turns out that humans win on the basis of guile or, you know, trickery of somehow, we're doomed.
Mr. GUTKIND: We are doomed. That's true.
(Soundbite of laughter)
CONAN: Let's see if we can get another...
Mr. GUTKIND: But...
CONAN: Oh, go on, I'm sorry. Go ahead.
Mr. GUTKIND: Well, just one other thing. This next step - we're now thinking robots - we're now teaching robots how to kind of think in our own shadow and with our own image. But the next step forward, what happens after we teach robots what we know and teach robots how to think? And that is we're beginning to teach robots learning algorithms. We're beginning to teach robots how to teach themselves, and how to learn themselves.
And when we do that - and we are, in fact, doing it right now - the human being, the scientist, the engineer, is losing a great deal of control over what is going to happen to that robot next.
CONAN: Let's go to David. David calling us from Reno, in Nevada.
DAVID (Caller): Hi, how are you doing?
DAVID: Yeah, I'd like to ask your guest, where do you try to balance or where do you draw the line with fuzzy logic - like applying fuzzy logic type of technology versus a true formulation of logic like black or white and make sure that one doesn't infringe upon the other?
Mr. GUTKIND: I'm not exactly sure what you're asking.
DAVID: Well, as far as being able to interpret using variables - as far as robots using fuzzy logic type of technology - how do you draw the line or where do you balance that versus using a program, say, that gives true formulation of either a right or a wrong or black and white type of logic?
Mr. GUTKIND: Well, this - the right or wrong type of logic is obviously the way good formulas and good algorithms begin. But if we're going to be able to teach our robots or make our robots much more autonomous, then we're going to have to give it the opportunity - give the robots the opportunity to think in a fuzzy way. To kind of - like human beings do.
DAVID: But how do you control at what point the fuzzy logic type of way infringes or crosses over into the, you know, the black or white type of logic? I mean, how do you control that?
Mr. GUTKIND: Why do we want to control it? I mean, writers…
DAVID: Because it isn't safe (unintelligible).
Mr. GUTKIND: …don't control what they're writing. It's - we don't narrow our efforts in a creative way. I think that this is - that writing for robots to think or to find their way is a very explosive and creative kind of maneuver and movement.
DAVID: OK. Well, thank you.
CONAN: Thanks for the call, David.
We're talking with Lee Gutkind today. His most recent book is "Almost Human: Making Robots Think." If you'd like to join us, our phone number here in Washington is 800-989-8255 - 800-989-TALK. You can reach us by e-mail as well. The address is firstname.lastname@example.org. We also have a blog on our Web site, it's npr.org/blogofthenation.
You're listening to TALK OF THE NATION coming to you from NPR News.
And let's get Sam on the line - Sam with us from Danville, California.
SAM (Caller): Hi.
CONAN: Hi, Sam.
SAM: I'm an engineer. I worked in robotics in the very, very early stages of robotics. And we were first, you know - back when the Apple II was actually a popular computer. But us as engineers, we're so proud and we are so amazed at our accomplishments that we - really what we need to do is look at whether we should do some of the things that we're doing, because if we don't look closely at what we're doing, we're going to end up ruining ourselves.
CONAN: What are you afraid of, Sam?
SAM: What am I afraid of?
SAM: Well, I lost my job three times to robots. A lot of us are losing our jobs to robots. And if we create something that is autonomous - not to debate the whole God theory and everything - but the human strayed, what us makes us think that the robots aren't going to stray? And what makes us think that we really will remain in control of them?
CONAN: Lee Gutkind?
Mr. GUTKIND: We don't know exactly whether we're going to remain in control. I can just tell you that roboticists now are not trying to create creatures that will run away with the world. The whole idea with robotics is - the roboticists are not developing robots for the sake of having fun or developing these mechanical creatures.
There's a reason. There are many important reasons that robots can exist. And I named a couple, like working with elderly people and working in hospitals and being the pioneers on the moon and Mars.
And yes, it's probably true that in our greatest fear that robots may well take the knowledge from every human being in a surgical process or suck it out in some sort of mysterious vacuum and take over the world. But I don't think that that's going to happen. And none of us want it to happen.
I think what we're trying to do is bring together technology - the software part of technology in the case of robots, and the hardware part of technology -and make it serve humanity overall.
SAM: That's what the engineers are trying to do, but corporations want to eliminate the bottom line and eliminate costs, and they do that by replacing people with machines. Now if you take away a nurse that has a heart and cares about a human being that they are taking care of in private situation and you put a machine in there, is that machine going to comfort that person?
Mr. GUTKIND: I don't think we're taking away a nurse, and I don't think that a robot very soon will be able to take away a job from a nurse. But robots can do, as I said to you, some really important things, really essential, but necessarily important things, like say deliver medications so that a human being, a nurse, can pay more attention to the patient and not be distracted by jobs that a machine can do.
CONAN: And let me follow up on what Sam's talking about. Lee Gutkind, are the engineers that you talk to and spend time with, and the computer scientists, are they thinking about some of the points that Sam raises?
Mr. GUTKIND: Well, I'll tell you. There is this tension, and has been for a long time, between the engineers - because the engineers really moved the robotics movement forward initially, and slowly but surely the software folks, the software geeks, have been catching up and taking the power and the influence away from the engineers in the robotics world.
And I think that there is this hardware/software rift, this resentment generally in the world - in the engineering world, because they have kind of lost their place in the robotics world. Right now, robotics is pretty much controlled by people who are writing code and, in fact, making robots think.
Someday soon, this code and the hardware that operates the code will come together again and work together more smoothly. But the engineers are, you know, in a much less influential place than they once happened to be in the world of robotics.
CONAN: Sam, thanks very much for the call. We appreciate it.
SAM: Let's just keep our hearts.
CONAN: Lee Gutkind, thank you for your time today. Thanks for coming in to join us.
Mr. GUTKIND: Terrific to talk with you.
CONAN: Lee Gutkind is the author of "Almost Human: Making Robots Think." He's the founder and editor of the literary journal "Creative Nonfiction," and he joined us today from WFAE, our member station in Charlotte, North Carolina.
When we come back from a short break, an argument that we are living in perhaps the most peaceful time in history. It's the TALK OF THE NATION after this break.