Interview with Kyujin Cho

Q1: You mentioned your dad was a pilot and that got you interested in robotics or mechanical engineering. Can you maybe start from the beginning?
Kyujin Cho: So, yes. Originally, I wanted to be a pilot. I lived in airbases, so I saw a lot of planes flying by and wanted to fly planes. But then when I became a fourth grader, my eyesight became bad. If your eyesight is bad, you cannot become a pilot. So I wanted to build planes. That was my original goal. But then when I was in high school, my high school teacher said, “You wanted to go to aerospace engineering, but if you go to mechanical engineering, you can build planes, you can build cars, you can build robots. If you go to aerospace engineering, you can only build planes.” So I went to mechanical engineering.
In my second year, we had a class where we built robots. The robotics class was called Creative Engineering Design. This was back in 1993. At that time, we didn’t have makerspaces or any way to experience building those kinds of robots. The only thing I had was that I liked building toys and plastic models. That class was actually started by Professor Chu, who was a MIT alumni and a professor at Purdue before coming to us. He started this class in collaboration with MIT. MIT 2.007 is very similar to that class. It was the first time that class was offered, and I was lucky to be in that class. I really enjoyed it. I was also picked to be in the final list where we were sent to Japan, Tokyo, for two weeks. We built robots with students from MIT, Cambridge, Darmstadt, Tokyo Institute of Technology, and São Paulo University. It was a great experience. Our team won the competition. I think that changed my life. That’s when I started thinking about a career in robotics. Although at the time I was still not sure what I was going to do after I graduated. I wasn’t even sure if I wanted to study more or whether I was going to go to a company.
But through my undergraduate studies, I took more project-based classes where I had to build things, and I found out that I really enjoyed those kinds of classes and challenges. So I decided to go abroad to study. Still, I wasn’t sure I was going to study robotics because back in the 1990s, robotics was not as popular. There wasn’t much funding in robotics at the time. Professor Frank Park took a class and said robotics was a very poor research area at the time, just like AI was for a long time. So I wasn’t sure if I could major in robotics. I went to MIT thinking I wanted to do manufacturing machines. But when I went to MIT, there was no funding in manufacturing either. All the professors I met said, “There isn’t much funding.” So I said, “Okay, I wanted to do robotics and mechatronics.” I went to Professor Hari Asada, who became my advisor. At the time, I had a fellowship for one semester, so he accepted me for one semester. That’s how I began my real research into robotics. Originally, Professor Asada was also working in healthcare and sensors because there wasn’t much funding in robotics. So I started working in the healthcare sector, creating things like wireless stethoscopes. Then around 2002, I was able to work on a robotics project. At the time, we were using artificial muscles to create new types of robots. That was Professor Asada’s research interest.I started working on how we could use artificial muscles. If you start using artificial muscles, you have many actuators and many degrees of freedom. So how do we deal with those kinds of high-degree-of-freedom systems? That became my PhD thesis: using synergies to control robotic hands.
After graduating, I wanted to do a postdoc. I was lucky because that was around the time when Robert Wood came to Harvard. He was looking for a postdoc, and I became the first postdoc in Robert’s lab. I think I was pretty lucky. At the same time, it was difficult because I had to establish the lab. There wasn’t much equipment, so we had to work with the equipment we already had without buying new equipment. That was a hard time, but I really enjoyed the new concept of bio-inspired robotics. That was also the time when soft robotics was starting to begin. This was around 2007 or 2008. Although soft robotics itself wasn’t really big at the time, there was the Campbell project, and I was working on artificial muscle actuators to create micro fish robots. My interest grew in creating new types of robots using new types of materials with unconventional designs. That’s how I started working on bio-robotics and soft robotics.
Q2: You mentioned that in the 1990s there wasn’t much funding. What kind of robots were you building in Japan in your classrooms?
Kyujin Cho: At the time, it was just a simple robot with two wheels and a Portland crane-style arm. Our project was to pick up cans and balls. We had a platform, a stage, and on the stage there were cans and balls. We were supposed to pick them up and put them into a basket. That was the kind of robot we built.
Q3: And do you remember the first robot you saw in person?
Kyujin Cho: The first robot I saw in person? That’s a good question. If I could call the robot that I built a robot, I think that’s the first robot I actually saw. Of course, I saw humanoid robots in the 1980s and 1990s. There were things like humanoid robots on display when we were young. I’m not sure if you know a robot called Mazinger Z, Masinga, or Taekwon V. Taekwon V is a Korean martial arts robot. When you say Taekwon V, it’s a humanoid robot that can do Taekwondo. It’s a human-like robot that can kick and punch and fight against evil forces. That was a very popular animation. If you went to some parks, there were huge Taekwon V robots showing movements. I guess that was the first robot-like thing I saw. I’m not sure if I can really call that a robot. It was more like a mannequin-style robot that moved slightly. I guess the first real robot I saw would be the robots that I built and the robots my friends built. Maybe you cannot even call that a real robot, but to me, that was a robot. It was a tele-operated robot with controllers, and we were supposed to complete missions. It was a robot that could actually do real missions. I guess that’s how I began robotics as a mechanism designer. At the time, everyone was building their own mechanisms. Some had three wheels, some had four wheels, some had two wheels, and some had an arm. Everyone was building a different arm.
When I was in Japan for the two-week competition, all kinds of new ideas appeared. One team came up with the idea of blocking the basket with a balloon. They had an arm that came out toward the basket of the other team and blocked the area with a big balloon so we couldn’t score. They kept it a secret until the last day, the day before the competition. At the time, we had a hole at the top and a hole on the side. They were coming in from our top hole, so I said, “Let’s build a tower. We’ll block our top hole so they cannot enter our basket.” That robot was actually shown on NHK. The two-week competition was filmed by NHK and aired on Sunday at 6 p.m., a prime time slot, because our team won. Our team was on the show quite a lot, and there was a part where I came up with the idea of what we should do. I think that kind of experience—where I came up with an idea and realized it could create something useful—was the major experience I had about robotics from the beginning. To me, it was something I created with a new concept and new ideas, and then it served a purpose. That was robotics from the beginning.
Q4: And has that sort of research question or approach changed over time?
Kyujin Cho: Yes. Concept-wise, it was about creating new mechanisms. I was playing around with different concepts: bio-inspired mechanisms, soft robotic mechanisms, artificial muscles, and how to use artificial muscles in design. Those kinds of new concepts came in, but it was always about creating new types of robots using new mechanisms. To come up with that mechanism, you have to understand the problem, understand the basic properties of what you’re handling, and then come up with a new design. So I guess from the beginning, that’s how I began my robotics career, and that’s still what I’m doing right now.
Q5: And you’ve had this fabulous career in robotics from those competition days to today. Have you experienced any challenges in your career?
Kyujin Cho: Yes, a lot. Of course, a lot. When I moved on to graduate school, one of the hardest things for me was at MIT, people were creating totally crazy things. I wasn’t used to those kinds of concepts. When I was in Korea, we were very practical. When I went to MIT, I was thinking, “Should I build a semiconductor machine?” I was interested in mechatronics and building machines, so I knew I wanted to build machines even if they were not robots. I was thinking about building semiconductor equipment because the semiconductor business was becoming big. That was my mindset when I went to MIT. But no one was building that kind of traditional thing. Everyone was working on totally new ideas and new concepts. That was the first challenge I had because Korean students are used to solving really hard problems and difficult math problems, but we are not used to these kinds of unusual ideas.
Q6: Can you give us an example of the weirdest thing you saw?
Kyujin Cho: One of my colleagues was building a bed that could turn people over. One of my first projects was to build a chair that could move the surface. Professor Asada had a project where he said, “Let’s put an eggbeater under the wing of a duck. Ducks cannot fly, so let’s give it an eggbeater and make it fly.” These kinds of crazy ideas. At first, I thought, “Why do you need that? That doesn’t sound serious. That doesn’t look like something that will become reality.” But gradually, I understood that academia should work on these kinds of new concepts and ideas because industry cannot do that. Industry has to be focused because they have to earn money. They cannot just work on random ideas. But academia can do that. We are supposed to do that. So I accepted that academia should work on these kinds of ideas that people think are crazy. That is good research. Then I started working on those kinds of new types of robots. Another challenge came after returning to Korea. At that time, what I was working on became considered unusual. I was working on soft robots. These days soft robotics is a major field, but back in 2009, when I started working on soft robots, there wasn’t even the term “soft robotics.” We were building robots out of fabrics. We were using sewing machines back in 2009. People were building origami-based robots, folding paper to create structures. We actually used paper to create a wheel. I did that because the concept was important. I liked the concept that if you use origami, you can simplify the design. Instead of using many actuators to transform something, you can use a single actuator because origami is based on a single connected sheet. That concept was very intriguing to me. But if you showed it to people, they would ask, “What are you doing? Why are you building a toy?”
Q7: How did you convince your Korean colleagues that what you were doing was worth exploring?
Kyujin Cho: At the beginning, I didn’t really convince them. What was important was the belief that this was the right way. A lot of that confidence came from visitors from the US. I had Henry Christensen, Paolo Dario, and others visit our lab, and they really liked what we were doing even from the very early beginning. I have to thank Frank because they came to our university because Frank invited them. They came to my lab, saw what I was working on, and told Frank, “He’s working on something great.” Their encouragement made me confident in my direction. I also had to convince my students because they were not used to these concepts. One student who was building my soft robotic hand, the ExoGlove, asked me, “Professor, is this a robot? Can you call this a robot?” I said, “Yes, this is a robot.” I had to go through the same process with my students that I went through myself. I needed confidence, but looking back, it was the community. The community and the soft robotics community. I started working on soft robotics back in 2008, when it wasn’t a big community. It was actually a field that was sometimes looked down upon because people said, “Is this really a robot?” But we had the soft robotics community led by Cecilia and researchers from Europe. That community helped form what became the field of soft robotics. Being part of that community allowed me to keep working on what I believed in without being shaken by people saying, “You should work on this” or “You should work on that.” Our work on soft robotics actually began not from the concept of soft robotics, but from meeting a user. I was trying to build a robot to help a disabled person feed himself. We were building a robotic arm to feed a person who could not use his arms and hands. But this person said, “I don’t want a robot to feed me. I want to feed myself. If I could use my hand, I could feed myself.”
At the time, robotic exoskeletons were very large. They had frames. We asked, “Why is it so big?” The answer was because of the frame. Then we asked, “Why do we need another frame? We already have bones. Bones are frames.” That’s how we began our work on soft robots. Originally, we called it a frameless exoskeleton. We were not good at naming things because there was no ChatGPT at the time. We started building robots out of fabrics and attaching tendons. The good thing about working early in a field is that you can meet challenges before others do. We were able to define the issues early and come up with ideas like using fabrics and polymers to create a hand called the ExoGlove. That became quite popular around 2015. So that was a very meaningful experience.
Q8: To this day, what would you say by traditional standards is your most absurd project you’ve done?
Kyujin Cho: Most absurd project I’ve done? That’s a pretty good question because “absurd” could mean it’s absurd to me, but not absurd to others. Or it could mean absurd from an outsider’s point of view. There are a lot. Like I said, working on using fabric to build robots was one. Using paper to build robots was another. We also built a robot that could jump on water. We were trying to build a robot that could jump on water, and people were asking me, “Why are you building a robot that jumps on water?”
Q9: Why were you building a robot that jumps on water?
Kyujin Cho: Because it’s fantastic. It’s amazing. If you can jump on water, how great is that? It’s the secret of nature. At the time, no one had analyzed how water striders can jump on water. So we first had to understand how water striders jump on water and then prove it using a robot. It turned out that our robot was the first robot that could jump on water. Then we understood what was really happening. Isn’t that interesting? It’s amazing, right? We had a sort of a hunch: maybe if we do this, we might be able to jump over water.
Q10: Do you have that video somewhere?
Kyujin Cho: Yes, yes, I will give you the videos. One of the reasons we could tackle those kinds of problems was that I told my students, “You don’t have to succeed. Even if you don’t succeed, you are going to be solving a problem that no one has ever looked into.” That experience will be a fundamental experience for you. You will be able to tackle unknown problems because to tackle unknown problems, you need experience doing that. If you didn’t have that kind of experience in school, then later in industry or outside of university, you’re not going to be able to do that. I wasn’t just hoping to solve the problem. Of course, I wanted to solve it. But I told my student, “Even if you achieve 50%, that’s fine.” But then we solved it. We found and explained how water striders jump on water, and we built a robot that could actually jump on water. That was a fantastic experience for us. But probably for other people, it seemed absurd. They asked, “Why are you working on this?” But after it was published in Science, a lot of media became interested. No one could ask, “Why are you working on this?” anymore. They were asking, “Why is this so fascinating?”
Q11: You mentioned a bit about naming things, but you were defining problems from the early days. How does this influx of generative AI and machine learning affect your research?
Kyujin Cho: I think the best thing for us is that our students don’t speak English as their first language, so it takes a long time to write papers. But now with AI, it has become much easier. It shortened the time needed to write papers. I guess that’s the best thing for us. In robotics, large language models have totally changed everything. Now people believe that humanoid robots can really work because of large language models and AI. I think that has changed how people view robotics. It has had a huge impact on how people think about robotics. Now people think robots can really do something, even with high degrees of freedom, like humanoid robots. That is a very positive change. But at the same time, we also have concerns that AI and large language models are not going to solve all robotics problems. Some people think large models will solve everything in robotics, but I don’t think they will solve all the problems.
Q12: What problems do you think AI cannot solve?
Kyujin Cho: Of course, AI will solve some problems. But for robots to be used in reality, they have to be successful 99.9% of the time, or even more. Especially now, we are saying robots will be used in everyday environments. Robots used to work in very structured environments, so they were safe. There were usually no people around, and robots just did what they were supposed to do. But now people are saying we are going to use robots in everyday life. That means robots will be more prone to failure. If a robot fails, it could be disastrous. Imagine you have a humanoid robot next to you and it falls on you. That could be dangerous. So it will take time. I hope people continue trying to solve those problems. Another concern is the huge amount of data needed to make this happen. We need a huge amount of GPUs, computation, energy, time, and effort to collect all the data. Even after collecting all that data, it is not perfect.
So how do we make this a reality? Even self-driving technology is still not perfect. It is still around level two or level three, even though technology is being developed toward level four or five. In reality, the systems that can actually be used still require human guidance. Can robots be used under human guidance? That defeats the purpose because robots are supposed to replace human work. If someone has to stay next to the robot all the time, that is what self-driving cars are doing right now. Even Waymo has people overseeing operations. So can robotics really do that? Even though there are high hopes, I’m also worried that if it does not become a reality and if it does not make money, what will happen if people become frustrated? I hope that doesn’t happen, but I am a little worried about that.
Q13: There is some anxiety among the public about robots replacing humans. Do you think that is the ultimate goal we are pursuing in robotics?
Kyujin Cho: Yes and no. We need robots to replace the jobs that people do not want to do. Now it has become very evident that in Korea and in the US, there are jobs where you cannot find people willing to work. You cannot find workers to do those jobs. So we need robots to do those kinds of jobs.
Q15: What are some of the jobs that people don’t want to work in?
Kyujin Cho: For example, construction work, factory work, and similar jobs. If you go to factories, most of the work people are doing is basically serving the machines. There are these large production lines, and people are there to put materials into the machines and take products out of the machines. Basically, they are helping the machines do their jobs. They are often very repetitive jobs. Some are fine, but some require a lot of physical force and have a higher risk of injury. Those are the kinds of jobs I think need to be replaced. I think all countries have become very prosperous. I don’t want my children to work in a factory. Even in China, I heard people from industry say that before, when labor costs increased, companies moved to China. After China, they moved to Vietnam and Indonesia. But now they are saying there is nowhere else to go. There is no cheap labor cost anywhere in the world anymore. So now we need robots. Whether they are humanoid robots or other types of robots, I think we definitely need robots. Think about the aging population in Korea. By 2040 or 2050, a large portion of our population will be elderly, with more than 40% of the population over the age of 60. Many people will start having difficulty walking and doing daily activities. They will not be able to get help from other people because everyone will be older. So we need robots to help. It could be humanoid robots, or it could be other types of robots. We need something that can help elderly people. Without that, people may have to stay at home or have real difficulty doing everyday activities.
We started working on robots for people with disabilities, but actually everyone becomes disabled before they die. So we believe these kinds of robots are needed. The technology is still not good enough, I think. I hope there is a big push in robotics, just like there is in humanoid robotics, because I think we all know robots—especially soft robots—are really needed. People want to be helped by robots, but they do not want to be seen being helped by robots. What people want and what people need sometimes do not match. As a researcher, I can try to build what I think people need, but if people do not want it because it is not what they want, or if it is difficult to use, then they will not use it. So we need to build robots that people actually want.
Q15: Where do you see your biggest contribution to robotics?
Kyujin Cho: My biggest contribution to robotics? I think our lab is good at designing new concepts and unconventional mechanisms. Starting with soft robots, bio-inspired mechanisms, and origami mechanisms, I see people following our work and exploring those areas. I think we expanded the concept of robot design. When you talk about design, traditionally there were certain designs people thought about. But we introduced new designs and expanded the range of things people could work on. I hope soft robotics becomes a really important technology one day because I think it can truly help people. If you have trouble with your knee, you stop walking. You stop going to the mountains or doing activities you used to enjoy. Then you start becoming frail. If we can help people walk without pain and maintain their mobility, I think it can make people healthier and happier. I think that would be a great achievement.
Q16: Do you work with industries?
Kyujin Cho: Yes, we do work with industries.
Q17: How do you approach them and convince them to fund your research?
Kyujin Cho: Actually, they approach us. I don’t really approach industry; they come to us.
Q18: What kind of industries?
Kyujin Cho: All sorts of industries. Electronics companies, flexible display companies, manufacturing industries, and even shipbuilding companies have come to us. They come to us because we show these kinds of novel ideas. These days, Korean industry is almost number one in the world in terms of technology. They need new technologies. Before, we were copying other people’s technology, so companies didn’t really need to look into what universities were doing. But now they need new ideas and new concepts to remain competitive. They see our work and think, “Okay, let’s try that.” Of course, it doesn’t mean it is immediately going to work, but it is something they can try. What we do is create a concept and show a prototype. Then companies can connect that idea with their own problems. They come to us and say, “We have this problem. Can you solve it using your technology?”
Q19: Do you actively engage with the media to put your work in public?
Kyujin Cho: I cannot say it is very active. Once we have a good paper, if we have a press release, that is how it gets picked up.
Q20: Why the press?
Kyujin Cho: Because then industry can see it. We have had several occasions like that. Sometimes we meet industry representatives and show them what we have. Then they say, “Can we use that to do this or do that?” For example, our object-gripping research was one of our recent works. Most grippers pick up objects one by one, one at a time. But humans can pick up several objects at once and move them. About five or six years ago, I saw a video of a worker picking up several objects and doing that. I thought, “Why can’t robots do that?” So we solved the problem using design. We came up with a multi-object gripper that could handle several objects. The concept behind our work is that we always try to find the principle behind a concept and then make use of that principle. We do not have to be confined to a certain shape or design. The principle is what is important. That is what we did with our water strider jumping robot, our soft robots, our multi-object gripper, and our origami-based wheels. We first come up with a principle and then build upon that principle. The multi-object gripper was also based on understanding the principle of how humans use their hands. Humans use their hands not only to pick up things but also to store things. When you pick up small objects, you store them on your palm and then pick up other objects. So we thought, “We need to divide the hand into a gripping portion and a storage portion.” Using that principle, we designed a gripper that has both a picking-up portion and a storage portion.