Introduction Excerpt for Feeling Machines

INTRODUCTION

Care Robotics 1.0

Someday.

Minoru Murata¹ sat on the edge of the bed in his living room, nibbling on a rice cracker he had just picked from the table of food in front of us. He leaned forward and raised an eyebrow.

“Sounds great. When can I get one?”

I couldn’t answer. I had no idea.

We were sitting together on a frigid January day not long after the New Year’s festivities of 2011 and only a few months into my sabbatical research on robotics technologies for eldercare in Japan. During this time, I lived on the second floor of the Murata home in western Tokyo in a room that Minoru and his wife, Keiko, both in their seventies, had rented to overseas researchers for years. We chatted occasionally as I came and went but we had precious few opportunities to talk. The slower pace of life around the new year, customarily a time for family gatherings, finally gave us a chance to get acquainted.

A few years before we met, Minoru suffered a cerebral aneurysm that left him with only partial use of the right side of his body. Since he could no longer manage the steep stairs that led to the second floor of the house, the couple relocated to the living space on the first floor, which they had originally rented to tenants. Now confined to a bed that had been moved into the living room, Minoru relied on his wife to help him with everyday tasks. His lack of mobility was a source of some frustration. Before the aneurysm, the Muratas had loved traveling abroad. They also enjoyed socializing with friends and colleagues, or with the foreign scholars they hosted. This was the worst thing, he would often say. He sorely missed the taste of beer and sake.

Encouraged by him nonetheless, I sheepishly took a sip of beer and sampled the many dishes Keiko had prepared for all of us. I looked up and met his gaze.

“It’s not ready yet. Unfortunately, you’ll have to wait.”

We were talking about my trip to a robotics manufacturer in neighboring Ibaraki Prefecture. I had recently visited the showroom of a company that makes the robot HAL, a wearable exoskeleton meant to help those with partial paralysis of the limbs regain some mobility. By means of strategically placed sensors that adhere to the surface of the wearer’s skin, the exoskeleton detects faint electrical traces of nerve signals as they travel to the extremities from the brain. These signals are processed in real time by a computer housed in a module on the waistband of the exoskeleton. The computer then triggers motorized joints on the hips and knees that, in combination with an additional system of sensors, support the movement intended by the person wearing it.

On my visit, a young associate in the showroom placed sensors on my right bicep and forearm. As I moved my arm up and down, part of the exoskeleton moved along with me. “Now watch this,” he said, smiling in anticipation of my response. I lifted my arm up again but this time he held my forearm and told me to keep pushing. To my surprise, the robotic arm continued upward as it had before. Holding my arm was an action meant to simulate a physical impairment.

It was a demonstration meant to amaze and amaze me it did. But it was only a demonstration. Despite the glitzy showroom in which it was presented, the HAL exoskeleton was not salable to consumers for personal use. It was available only for lease and at a cost out of the reach of most individual consumers. Additionally, at the time, its safety had yet to be scientifically verified. Rather than individual elderly, its target markets were institutions like hospitals and nursing homes. Only in such places could such a new device be closely monitored, studied, adjusted, and later improved.

Like so many robots I encountered that year, it was not yet ready.

Mr. Murata looked at me, agitated. “Well, tell them to hurry up and make something I can use!”

I nodded. I understood his frustration.

*   *   *

Elsewhere.

“If you really want to see how people use my robot, you should go to Denmark.”

In summer 2009, Dr. Takanori Shibata and I met in the lobby restaurant of an Osaka hotel. Shibata is a kinetic presence. He has a quick mind, speaks rapidly, and always seems in a hurry for another appointment. In fact, he usually does have another appointment. He is the inventor of Paro, a furry white robot that looks like a baby harp seal and responds to speech and touch. At the time that I met him, Shibata had been developing iterations of Paro for over a decade. Fluent in English, he often travels internationally to present research on the therapeutic benefits of his robot and to promote its adoption. I felt fortunate to have the opportunity to meet him in person.

His robot Paro was emblematic of just the kind of “care robot” (kaigo robotto) that I planned to study in Japan, and earliest contacts in the field all referred me to him. When we finally met, I told him about my interest in studying the use of care robots among older adults in Japan. He expressed support and said that he would be happy to provide referrals, but he cautioned that there were only a few nursing homes in Japan where Paro was regularly used. Instead, he told me, the vast majority of Paro sold in Japan were purchased by individual users for use as electronic pets. When I asked him if he could refer me to some of these users, he demurred, citing privacy concerns.

Instead, Dr. Shibata shifted the conversation quickly to the use of Paro in Denmark. A pilot project headed by an enterprising nurse in Copenhagen had concluded the previous year. The project tested Paro’s utility in the care of older adults with dementia. The success of the project persuaded the Danish government to order approximately one hundred Paro robots for nursing homes throughout the country. Shibata had recently returned from a visit and had witnessed firsthand how enthusiastically Paro had been taken up. He urged me to go see for myself.

I continued my work in Japan but I would take his advice.

*   *   *

Otherwise.

Run Jenny Run.

A few months before the Muratas and I had our meal together, Thomas Schildhauer, a trauma surgeon at Bergmannsheil Hospital in Germany, attended a meeting in Düsseldorf with the Japanese roboticist Yoshiyuki Sankai. Prof. Sankai is faculty in the Graduate School of Systems and Information Engineering at the University of Tsukuba and inventor of the exoskeleton HAL, the very device that I had described to Minoru. As he had on many other occasions, Sankai demonstrated how the sophisticated hardware and software undergirding the exoskeleton worked. Interspersed among the many slides detailing how the mechanical and electrical systems interact within the machine—a system philosophy Sankai calls “cybernics”—were demonstrations of the exoskeleton performed by one of his young, able-bodied assistants. Wearing the HAL suit on his lower body, the assistant strode across the stage without any visible encumbrance; to the contrary, the demonstration climaxed with the assistant lifting a pile of heavy weights without straining. The demonstration drove home Sankai’s point: his device could not only help people with disabilities walk independently but could also ease the physical strain of caring for the aged or physically impaired. The demand for such a device would only increase, Sankai added, as the world continued to age rapidly. This was a problem felt acutely both in Germany and in the most aged society on the planet, his home of Japan.

Sankai’s presentation was met with general skepticism. Most of the assembled bureaucrats and business managers doubted whether his machine could work in Germany. Dr. Schildhauer, however, thought otherwise. In his hospital work, he operated on patients with traumatic spinal cord injuries and oversaw their programs of post-surgery rehabilitation. He had found that getting a person to move a limb as much as possible, and as quickly as possible, after surgery helped that person recover the greatest amount of function. Might it not be worth trying out the HAL suit in the rehabilitative care of spinal cord injury patients who had recently undergone surgery?

The doctor approached Sankai after he had concluded his presentation and told him about his idea. Intrigued, Sankai invited the doctor to present his ideas at an international research conference in Japan the following spring. That presentation led to a pilot project on the use of HAL with spinal cord injury patients in Germany. The success of the pilot project led to an application to European Union authorities to certify HAL as safe for medical use. The granting of this certification led in turn to investment by Cyberdyne, Sankai’s venture firm in Japan, in a facility adjacent to Schildhauer’s hospital in Germany. The establishment of this facility elicited an additional round of funding from the New Energy and Industrial Technology Development Organization (NEDO), the research and development arm of Japan’s Ministry of Economy, Trade, and Industry, as well as a new research trial with HAL in Germany. That research trial required the recruitment of several participants.

One of them was an ebullient young woman, who later sat across a conference table from me, awaiting my questions. About eighteen months before I met her, Jenny had an accident that partially severed her spinal cord. The accident left her paraplegic and unable to walk without crutches. She underwent standard rehabilitative therapy for six months until she heard about the clinical trial with HAL. Jenny told me that she had made significant progress using the HAL exoskeleton and wished to continue working with it. However, the trial had finished and the therapy was not yet covered by the German healthcare system. Aware of these limitations, friends of Jenny organized a social media campaign to raise enough money for her to continue her rehabilitation. Her friends named the campaign Run Jenny Run, an allusion to the internationally successful German film Run Lola Run (1998, dir. Thomas Tykwer). Jenny’s actual aims were much more modest than this virtual campaign suggests—she didn’t believe that she would ever run again. Nevertheless, the initiative garnered enough support that Jenny was able to continue her therapy, at least for the time being. Unless insurance coverage were expanded to cover HAL “training” (as post-rehabilitative care is called in the facility), it was unclear how much longer she and many other patients in Germany would be able to access the robotic exoskeleton from Japan.

*   *   *

Each day seems to bring news of another new robot about to arrive in the world. As I began my early work on this book in summer 2019, for example, the website The Verge reported that the American company Boston Dynamics planned to launch Spot, a much-heralded quadrupedal robot that moves with close to the agility of a real dog, as a commercial product later in the year (Vincent 2019). That story was posted just hours after an online feature by the magazine Wired on a tech start-up engineering robots to deliver food via bike lanes and a Washington Post article on Tesla’s plans to convert its vehicles into fully autonomous self-driving cars—that is, a robot on wheels—by the end of the year (Davies 2019; Siddiqui 2019). And that was just one day. Extending the search to a week or several months would yield stories on robots that do surgery, robotic drones that surveil the skies, robots that deliver packages by air, robots that vacuum, robots that mow lawns, robots that cook, robots that work in warehouses, robots that harvest crops, robots that work in hospitals, robots that perform music, robots that tumble like gymnasts, among others. Some of these robots are already here; others remain promises for the future. Whether experimental or actual, the dynamism of contemporary robotics has sparked a range of popular writing on robots and automation, prompting one scholar to claim that we are living in a “robotic moment.”²

What does it mean to live in a robotic moment? For some, it means that the times are exciting, filled with the expectation that intelligent machines might help make our lives easier. For others, the moment is charged with fear. Robots, and their increasingly “smart” digital cousins, might grow so powerful that they wrest control from their creators to become malevolent overlords, enslaving humans in the way that humans once controlled them. For still others, robotics technologies imperil the economic security of workers, threatening to automate jobs away at a faster pace than new ones are created for those who are displaced. The robotic moment, then, seems to be as much about the future of humanity as it is about the role of machines in the present.

This has been true of robots, in fact, since the moment of their creation. The term robot first appeared in the play R.U.R. (Rossum’s Universal Robots) written by the Czech playwright Karel Čapek in 1920. In Čapek’s play, robots are not the jumble of sensors, chips, motors, and metal with which we are most familiar today. They are organic beings, artificial humans that lack emotions and are made to work on behalf of their human creators (Čapek 2004). In the play, robots work so efficiently that they ultimately replace human workers entirely, but a production “defect” leads one robot to gain awareness of their collective enslavement. He (the robot is gendered male) leads a worldwide revolt that results in the annihilation of all but a few humans. The vision of productive but ultimately lethal robots in Čapek’s play has had a lasting effect on the Western imagination. Robotic moments then and now often both portend futures of leisure and prophesy the end of labor or even humanity itself.

This book, by contrast, explores another robotic moment, one concerned more with the care of humanity than its pursuit of leisure or its extinction. Around the turn of the millennium, Japanese roboticists began converting technologies developed for factory production into machines for the care of older adults and people with disabilities. Unlike the robots in Čapek’s play and today’s factories, these robots have been engineered to sense, anticipate, and respond to changes in human affect and intention; they are feeling machines, designed to be felt and programmed to feel. Over the past decade, so-called care robots have moved slowly out of laboratories into everyday practices of care at home and abroad. This movement has been facilitated in part by governments interested in supporting shrinking populations of care workers without adding pressure on already strained welfare resources. But it is not just the fear of fiscal crisis at work. This robotic moment is undergirded by a transformation in imagining the ends of care such that caring for the aged and disabled by/with feeling machines has become for some not only thinkable but desirable. This book examines the implications of this shift for our understandings of care and our evolving relationship with digital technology.

For scholars of Japan, it seems nearly inevitable that this robotic moment would arrive. Japan is a famously “aged society,” and has been for a long time. Twenty-eight percent of Japan’s population is currently over sixty-five, giving it the largest proportion of elderly of any country in the world (World Bank 2024). Japan also has one of the lowest birthrates among the low-birthrate countries of the Global North. The close relationship between low fertility and high longevity is widely recognized in Japan, so much so that the social issues are collapsed together as the “low fertility, aging society” (shōshi kōreika shakai). The aging society broke into public consciousness in 1989 and has only grown in intensity since then.³ In fact, a Pew Research survey found in 2014 that a whopping 87 percent of the Japanese public felt that “aging is a problem,” the highest of any country surveyed (Pew Research Center 2014). Over all this time, the country has maintained relatively strict limits on immigration. Only recently has the national government signaled greater openness to foreign workers and reduced obstacles to their establishing permanent residency, in the hope of partially alleviating a shortage of workers in an “aging society” (Smith 2019).

Given a predicted increase in the number of seniors, a projected shortfall of care workers, and continued low levels of immigration, when I began a year of fieldwork in 2010 there seemed few options for Japan other than to supplement human labor with new kinds of machines. Such a turn to machines was not without precedent; robots had help “save” the country once before. Industrial robotics took off in Japan after the oil shocks of 1973 slowed the economic boom of the 1960s, encouraging a shift toward the automation of factory work. The number of industrial robots in Japan rose every year from 1975 through the “bubble” period of the 1980s, boosting productivity and helping Japan reemerge as a global economic superpower (Schodt 1988, 115–20). For most of the twentieth century, Japan led the world in applying robotics technology in manufacturing, giving it a wealth of surplus intellectual and physical capital. An association between technological prowess and economic vitality cemented over this time, and many in Japan came to think of their nation as a “robot kingdom” (robotto taikoku; see Schodt 1988, 14). So, too, did contemporaneous observers outside Japan. In an edited volume on robotics, two American experts wrote that “the Japanese . . . are now the most advanced in robotics” (Engelberger 1985, 190) and that “in Japan . . . the robot revolution is further advanced than in any other country” (Wolkomir 1985, 233). Japan’s leadership in industrial robotics was matched by pathbreaking research on humanoid robotics and, later, in studies of human responses to lifelike androids (Brooks 2002; Frumer 2018a; Robertson 2018, 110–20). By the late 1990s, Japanese companies had pioneered a range of consumer products that built on this technological heritage, including Sony’s robot dog AIBO and innovative devices like Tamagotchi that newly engaged with consumers by eliciting affective connections and generating a sense of intimacy (Allison 2006).

Of course, what seems inevitable might not actually be so. When I began my fieldwork in Japan, the actual application of robots in eldercare was quite limited. I did meet individuals who were committed to using robots therapeutically among elderly adults with dementia. However, I encountered just as many roboticists who, though keen to share the promising results of short-term field studies, deferred the real-world application of their machines to the future. Demand for their devices, they assured me, would arrive someday, a day that would appear with the many older adults of a future “aging society.” These deferrals reflected the temporal inequities embedded in contemporary robotics; the sentiments of those who could afford to luxuriate in the anticipation of a technological future and those older adults, like Minoru Murata, who could not.

Still, while some oriented my gaze forward to a future of care robotics, others like Dr. Shibata and Prof. Sankai directed it elsewhere to locations where Japanese machines had already been embraced in the present. Shibata’s excitement about the adoption of Paro in Denmark made clear early on that Japanese care robots were more than just a local solution for a local problem of population aging. My study needed to move beyond Japan, but the incipient transformation of HAL otherwise into a tool for rehabilitation meant that it had to expand in scope, too, to include younger populations coping with traumatic injury. A project that I had expected to focus narrowly on human–robot interaction among older adults in Japan similarly grew otherwise into a study of a globally dispersed roboscape I call “care robotics 1.0.”

Notes

1. In this book, I use pseudonyms to refer to all individuals who are not public figures.

2. For mass-market writing in English on robots and related digital technologies like AI, see Brynjolfsson and McAfee 2014; Ford 2015; Hornyak 2006; Markoff 2015; McAfee and Brynjolfsson 2017; White 2015. On the “robotic moment,” see Turkle 2011.

3. In 1989, the total fertility rate dropped to 1.53. Fertility rates had been dropping for some time, but in this year rates fell below an aberrantly low rate in 1966.