Masahiro Mori at his home in Tokyo.
In this guest post, Norri Kageki interviews Masahiro Mori, who, as a professor of engineering at Tokyo Institute of Technology in the 1970s, proposed the now-famous concept of the uncanny valley. [Read the first authorized translation of his seminal article here.] Mori's insight was that people would react with revulsion to humanlike robots, whose appearance resembled, but did not quite replicate, that of a real human. He called this phenomenon bukimi no tani (the term "uncanny valley" first appeared in the 1978 book Robots: Fact, Fiction, and Prediction, written by Jasia Reichardt).
The uncanny valley has become more relevant in the past few years since robots that actually look and move like humans are starting to become a reality. In fact, researchers currently debate over whether they should try to overcome the uncanny valley or simply design robots that are more mechanical in appearance. It seems like a good time to hear again from Mori, who's now 85 and still full of ideas. Kageki, a Silicon Valley-based writer, met Mori at his Tokyo residence last November and conducted the following interview.
Norri Kageki: Your essay titled "The Uncanny Valley" first appeared in a 1970 issue of Energy (pictured below), a magazine published by Esso, a Japanese subsidiary of Standard Oil Co. How did this come about?
Masahiro Mori: Back then, an editor at Energy informed me that they were going to publish an issue titled "Robotics and Thought." He asked me to be a part of a round table on the subject along with the science fiction writer Sakyo Komatsu and Prof. Natsuhiko Yoshida (of Tokyo Institute of Technology), whose expertise was in philosophy and logic. I was asked to write something for the issue as well. Since I was a child, I have never liked looking at wax figures. They looked somewhat creepy to me. At that time, electronic prosthetic hands were being developed, and they triggered in me the same kind of sensation. These experiences had made me start thinking about robots in general, which led me to write that essay. The uncanny valley was my intuition. It was one of my ideas.
NK: I don't think there were any robots that looked like humans back in 1970. And yet you had this idea.
MM: That's right. There were no such robots yet. In those days, people didn't think universities should be doing research on robots. They thought that it was frivolous to be working on a "toy." As with anything new, there was much opposition, and I felt that everyone was against me. In 1970, it was too embarrassing for anyone to make a grant proposal to the Ministry of Education to start a project on robots. Even the students were dubious about it. They were concerned about their future careers after working on a project like this. Today, kids go to college specifically to study robotics. How times have changed!
The cover of the 1970 issue of Energy in which Mori's uncanny valley essay first appeared.
NK: There is a debate on whether the uncanny valley is a scientific concept or not. What is your view on this?
MM: I have read that there is scientific evidence that the uncanny valley does indeed exist; for example, by measuring brain waves scientists have found evidence of that. I do appreciate the fact that research is being conducted in this area, but from my point of view, I think that the brain waves act that way because we feel eerie. It still doesn't explain why we feel eerie to begin with. The uncanny valley relates to various disciplines, including philosophy, psychology, and design, and that is why I think it has generated so much interest. However, I never imagined that it would gain such a magnitude of interest when I first wrote about it. Pointing out the existence of the uncanny valley was more of a piece of advice from me to people who design robots rather than a scientific statement.
NK: How much response did you get when you first wrote your essay?
MM: You could say that there was no response.
NK: Then when did the uncanny valley start to receive attention from outside of Japan?
MM: It started to be picked up after the humanoids conference in 2005 (IEEE Robotics and Automation Society International Conference on Humanoid Robots).1 I was invited to that event, but I was unable to attend due to other commitments.
NK: Since then, the uncanny valley has become a sort of principle in the robotics field. What is your take on this?
MM: It is nothing for me to brag about, but I am pleasantly surprised that it has become so well known that people are talking about it so much.
NK: It means that it took 40 years for the world to catch up.
MM: Yes, I agree. This may not get across to people overseas, but I think of myself as the dog in the story "Hanasaka Jiisan." [In this Japanese folk story, a dog barks to let his owners know where to dig, and when they dig they find gold.] I seem to have a good nose for sniffing out interesting things, but I don't have the skill to dig them up. That's why I bark "Dig here!" and then other people will dig and find treasure. Or, if people don't dig, I dig a little bit by myself. Then people come in like a gold rush. The uncanny valley was just one of the things that I sensed.
NK: Do you think there are robots that have crossed the uncanny valley?
MM: Yes. I think the HRP-4C [an adult-size humanoid robot developed by Japan's National Institute of Advanced Industrial Science and Technology] is one of them. Though on second thought, it may still have a bit of eeriness in it.
Mori's now-famous uncanny valley chart.
NK: Your uncanny valley chart captures the concept nicely: The curve first goes up, as people's affinity toward robots increases, they become more humanlike, but only up to a point, when the curve suddenly plunges into the uncanny valley. Do you still think that robot designers should aim for the first peak instead of aiming beyond the valley?
MM: Yes, I do. I always tell them to stop there. Why do you have to take the risk and try to get closer to the other side? It's not even interesting to develop a robot that looks exactly like human, from my perspective.
NK: What do you mean by that?
MM: I have no motivation to build a robot that resides on the other side of the valley. I think the [design of] Asimo is more invigorating. I feel that robots should be different from human beings.
NK: But aren't there many projects in Japan that are aiming for the other side of the valley?
MM: I wouldn't say that there are many, but yes there are some, and that's fine. Although I do think it is difficult. Using the woodcarving of a Buddha statue as an example, that one last touch of the knife may destroy the whole thing. There is a narrow margin for error.
NK: There are people like former Massachusetts Institute of Technology professor and founder of start-up Heartland Robotics Rodney Brooks who think humans are also machines and that they can be built. What are your thoughts on this?
MM: I don't know. I do agree that some aspects of humans are like machines. But what I don't know is the mind. I think this is a problem that humans will never be able to solve. No one can explain whether an object can bear a mind. We don't know whether a mind will be formed when computers become really precise. Will a computer feel one day that it's not in a good mood or that it doesn't like this person or that? I don't know. My friend Ichiro Kato [the late Waseda University professor and pioneer in humanoid research] used to think that the mind can be created, but I don't think so. That's because we really don't know what the mind essentially is.
The uncanny valley relates to various disciplines, including philosophy, psychology, and design, and that is why I think it has generated so much interest.
NK: How would you like the idea of the uncanny valley to be applied in the world?
MM: Research related to the uncanny valley will make progress naturally. I am grateful for that. It has been 24 years since I retired from the university, and I am now 85 years old. There is nothing that I can do about it but to look forward to this field being cultivated by other people's abilities.
NK: In what areas do you think humanoids would be most useful?
MM: There are many things that humans can't do for which robots would be helpful, but I don't think the robots need to be humanoids. I actually think that it would be extremely difficult for a humanoid to help with housework. It's not possible to build one at the price points of vacuum cleaners and dishwashers.
NK: What is your definition of a robot?
MM: You can't define a robot. It's the same as trying to define Mt. Fuji. If a steep hill suddenly protrudes from the flatland, you can draw a line to show where the mountain starts, but Mt. Fuji becomes higher so gradually that you can't draw a line. Robots are like Mt. Fuji. It's hard to separate what is a robot from what is not. Asimo is so near the peak, anyone can easily call it a robot. But what about a dishwasher? It can automatically wash dishes, so you might call it a robot. The line is blurry.
NK: Are there any robots that you are interested in nowadays?
MM: I have lost interest in robotics hardware. I can see that the capabilities are improving, but there is nothing out there even now that I have not imagined before. For example, swarm robots. People now call them distributed autonomous systems, and there has been much progress in terms of their applications in the past ten years. I had thought about it many years ago, and I built seven robots that moved as a swarm and exhibited them at Expo '75 held in Okinawa in 1975. I think those were the first swarm robots in the world. Let me show you a sketchbook that I have kept [photo, below]. I used to cut out these pictures and photographs from various magazines, and there are designs that my students and I drew back in those days. These are from the 1960s before Xerox photocopies became available in Japan. I categorized them by fingers and shoulders [as references for designs]. I feel that people are finally building the robots that are in this sketchbook. Back then, I couldn't make these robots because I didn't have the funding. Now, my former students and their students are building various robots. For example, the Asimo and HRP-4C projects are being led by my former students.
Mori's sketchbooks with his robot designs.
NK: If you were still at the university, are there any robots that you would like to build? Is there anything you feel that you have left undone?
MM: How should I put this... I have never approached my research in a way where you decide on an area and then you try to dig out everything in that area. I walked around and when there was something that smelled interesting, I would bark "Dig here!" And then when someone else started digging there, I would move on to somewhere else. That's what I have been doing. So there is no area of specialty for me. Usually, academics will decide on a topic and then you can find all the papers on that topic at that person's laboratory and so on. I did work that way until I finished my dissertation but that way of research did not suit me well. So I roamed around and jumped on to whatever smelled interesting.
NK: Have you come across any interesting scents recently?
MM: Right now, I am working hard on the relationship between technology and the teachings of Buddha. To develop robots, you need to understand humans. I think the teachings of Buddha is the best way to understand humans, especially with regard to understanding the human mind. Technology is not all good, and in certain cases, the negative aspects of technology will appear strongly. We need to save technology from becoming negative. That is what's on my mind right now.
Norri Kageki is a journalist who writes about robots. She is originally from Tokyo and currently lives in the San Francisco Bay Area. She is the publisher of GetRobo and also writes for various publications in the U.S. and Japan.
Photos: Norri Kageki