SCOTT SIMON, HOST:
Robots are on the march - more accurately, humanoid robots that are designed to resemble human beings. The World Robot Conference is underway in Beijing. And among the presentations, there are humanoids that - or who - can dance, box - one that even plays ping-pong. But are these robots really about to revolutionize the world?
Naomi Fitter is a professor in the School of Mechanical, Industrial, and Manufacturing Engineering at Oregon State University in Corvallis. Thanks so much for being with us.
NAOMI FITTER: Thanks for having me.
SIMON: I feel the need for a primer. What exactly is a humanoid robot?
FITTER: I think your definition was pretty spot on. It's a robot that tends to be between infant size and, like, a full-grown adult size and height. And it has a lot of different features similar to a human being - things like legs, arms, often even faces and things like that.
SIMON: Why do you think we're seeing a growth in humanoid robots now?
FITTER: There are a couple reasons. One is that this idea of generalist robots that can do a broad range of tasks, I think, is appealing for many companies, many venture capitalists, even roboticists. And if you have a generalist robot instead of, like, buying a different robot for every task, you could potentially apply that robot for a broad range of activities. I think there's also appeal to designing robots in our own image. So perhaps even to some extent, there's interest in humanoids for that reason.
SIMON: What are some of the challenges about developing them?
FITTER: One challenge with humanoids - they're just so complicated. Typically, they have more degrees of freedom or things like motors in their construction than things like wheeled robots that drive around on the ground, more similar to a car. Another interesting thing with humanoids is they're more and more deployed in settings with day-to-day people who are just navigating, going about their business and maybe didn't anticipate coming across a humanoid when they were going about their day. So I think there are new, interesting human-robot interaction challenges, too, that will come about as more of these humanoid robots are in day-to-day settings.
SIMON: May I ask you - I find myself torn between being charmed and a little creeped out by humanoid robots.
FITTER: That's fair. We're studying this a little bit in our lab. We do see a effect a little bit like that in things like delivery robots, perceptions of those by the public. We see that, at first, there's often excitement and a little bit of a novelty effect, but that, as time goes on, there seems to be a bit of a sag in positive opinion. Another thing we've seen - we deploy some humanoid comedy robots, too. So we've seen kind of a mix similar to what you're alluding to there as well. People seem to like it when the robots are reading the room and giving responses to the crowd reactions that are spot on. Sometimes...
SIMON: I'm sorry. Humanoid comedy robots? I have missed this.
FITTER: Oh, yeah.
SIMON: Like a Robin Williams robot.
FITTER: Well, the one we use is a smaller, more, like, infant-sized robot. It's called the NAO robot, which has been out there on the market for a little over 15 years, so it's probably one of the longer-term humanoid robots that's out there. But we use the robot in comedy, and it's able to scan the crowd, read the room and try to work the room with its comedy. Some people enjoy that the robot can understand the scene around it and respond in real time to their reactions, and some people are kind of creeped out by it.
SIMON: Is China ahead of the United States in this field?
FITTER: It's a complicated question. I see a lot of new robot hardware platforms out of China. One thing with robot hardware is it really matters what processing, what kind of common sense and what software architecture is happening under the hood, as far as what the robot can do.
So a few years ago, at the IROS conference - one of the big international robotics conferences - Disney even did a demonstration to this effect, taking, like, relatively low-cost hardware components and making a robot that seemed extremely lifelike and animate. So you can do it with good-enough processing, with interesting-enough models, with a lot of learning and simulation, which then translates to robotic hardware. So it's a complicated question to answer, partly because it's not just the hardware that matters but also the processing, the intelligence that you're putting into the system.
SIMON: This summer, the FCC announced a ban on foreign-made robots, describing them as national security risks. What do you think?
FITTER: I think one important thing to note there is just when you create a robot, you could potentially leave backdoors in kind of the softer architecture of the robot that can lead to risks. So I understand some of the FCC reasoning there. We have seen that with some robotic platforms in the past. And the potential to kind of more easily hack into robots, to access them through these kind of backdoor means - it's a reason to have some degree of, you know, caution in mind. It could happen potentially with any robot. But it's something, for sure, to keep an eye out for.
SIMON: Naomi Fitter from Oregon State University. Thanks so much for being with us.
FITTER: Thanks for having me. Transcript provided by NPR, Copyright NPR.
NPR transcripts are created on a rush deadline by an NPR contractor. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.
