[Takanori Komatsu is a cognitive scientist at Meiji University in Tokyo who has been studying human-robot interaction by focusing on human reactions to robots with different characteristics and that take on different roles. Among other interesting elements of his work is his use of thought experiments like the Trolley Problem to understand our perceptions of and responses to robots. Medium-as-social-actor presence isn’t mentioned explicitly but is central to the ideas in the story below about Professor Komatsu, which is from Scientific American Custom Media; the original version includes a second image and an 8:51 minute video (also available on YouTube). –Matthew]

[Image: Takanori Komatsu is a cognitive scientist at the School of Interdisciplinary Mathematical Sciences of Meiji University in Tokyo. Credit: Meiji University]

Friend or Fridge? What Humans Really Want from Robots

Robots are increasingly capable, but they could do more in everyday life if we felt less uneasy around them argues a cognitive scientist in Japan.

November 30, 2023
This article was produced for Meiji University by Scientific American Custom Media, a division separate from the magazine’s board of editors.

When Astro Boy debuted 60 years ago in Japan, the cartoon about a robot boy with human emotions captured the imagination of a whole generation, inspiring many to become engineers. Since then, robots have leapt off the screens and into factory lines and operating rooms, particularly in Japan where there are 390 robots for every 10,000 workers, one of the highest rates per capita in the world.

But the situation looks different in everyday life. While we benefit from smart appliances such as automated vacuums and voice-controlled devices in our homes, robots aren’t replacing our friends or even pets any time soon. Despite decades of technological advancement, certain roles remain out of their reach.

“It’s not a matter of technology,” says Takanori Komatsu, a cognitive scientist at the School of Interdisciplinary Mathematical Sciences of Meiji University in Tokyo. “There’s a possibility we unconsciously feel uneasy about robots, and therefore refuse to make friends with them.”

ROBOTIC LIMITATIONS

While traditional research in human-robot interactions focuses on using engineering to give robots human-like movements, an alternative cognitive approach focused on human psychology has been gaining momentum since the mid-2000s. A mechanical engineer by training, Komatsu is now aiming to uncover how humans truly feel about robots — and why.

A one-time fan of Formula 1, Komatsu originally wanted to become an engineer of racing-car engines, but his interest shifted to robotics as an undergraduate student. A turning point came when he was a guest researcher at the Technical University of Vienna. While there, he encountered an event where a robotic vision system couldn’t capture the image of an object correctly when the amount of light reduced as day turned to night. “Humans wouldn’t face such a difficulty. I realized it’s very important to learn more about human’s ability and intellectual flexibility for robotic engineering,” Komatsu recalls.

As he pursued his research career in human-robot interactions, Komatsu found there had been little progress in the field, with similar arguments reiterated over many years. “I thought perhaps something crucial was being left out. Otherwise, how could we explain why robots have yet to be part of everyday life despite the hard work of so many capable researchers? Then it occurred to me that humans themselves may not be accepting robots.”

THE TROLLEY PROBLEM WITH A TWIST

To delve into the depths of the human mind, Komatsu turned from experiments with lab equipment to thought experiments. He was particularly inspired by the research of Bertram Malle, a psychologist at Brown University in the United States, which looks at the classic trolley problem; a runaway train is about to kill four workers on the track.

Someone standing by the tracks could save them by pulling a lever to switch the train on to another track, but that would come at the expense of another worker’s life. Malle was interested in what if the person making the decision was not a person at all, but a robot. He found that participants tended to blame the robot more for allowing four people to die through its inaction, suggesting that humans have different feelings toward the morals of robots. Thought experiments such as the trolley problem help to uncover the mental processes behind attempts to resolve moral dilemmas.

To see if such blame judgements are universal or vary culturally, Komatsu and Malle conducted a similar study to compare the responses to the trolley problem in Japan and the US. The team found Japanese participants were more willing to accept robots as decision-making agents in a hypothetical situation, while many people in the US were more skeptical about a robot’s abilities to do so. However, when it came to the actual decision to do nothing, people in both countries blamed the robot more than the human decision maker.

“The first outcome had been predicted because of Japan’s high affinity with robots,” Komatsu says. “But interestingly, the study also showed there are no cultural differences in moral judgments between the US and Japan. Eventually, I’ve come to the conclusion that Japanese people don’t have any special feelings toward robots.”

A WORLD WITHOUT (VISIBLE) ROBOTS

While Komatsu is keen to build evidence to support his claims by implementing a few more thought experiments, his findings already give us a clue to how people can better co-exist with robots. For example, Komatsu’s recent study with Toyota Research Institute shows that many people, especially those over 50, are willing to have robots as conversation partners to alleviate loneliness.

On the flipside, Komatsu warns that engineering efforts to give robots human-like personalities could lead the research in the wrong direction. “As robots look more like humans, people expect them to behave like humans and they become disappointed if robots fail to meet their expectations,” says Komatsu, adding that robots are more likely to be accepted if people see them as machines. “Robots in the form of smart speakers, for example, could be a great ally.”

A more fundamental issue, according to Komatsu’s earlier research, is that people perceive robots as ambiguous entities that are neither things nor humans. “The word ‘robot’ itself evokes an eerie machine-driven thing,” Komatsu says. He anticipates that robots will be gradually incorporated into social infrastructure such that people won’t notice they are there. The word robot — coined in a Czech novel about a century ago — might eventually disappear.

Still, Komatsu thinks it is important for engineers to keep developing robots because such efforts could lead to useful new technologies. In addition, inter-disciplinary collaborations spanning literature, philosophy and other fields unrelated to engineering and cognitive science at a glance, are vital to spur discussions on better interactions between humans and robots.

“Many people say I am pessimistic or uninspired, but I hope I can address important insights about how people perceive robots and make an impact on robot engineering,” he adds.