Read an Excerpt
CHAPTER 1
Robot Visions
Our dream is to create a society where it is nothing special for people to live together with robots.
Akifumi Tamaoki
BEGINNINGS
For much of my childhood in Kodaira, a suburban community west of Tokyo, I had to watch television at our neighbors' homes. We only acquired a black-and-white set in 1964 to watch the Summer Olympics, by which time the fields and chestnut orchards surrounding our house were being razed by tract-home developers and the gravel roads paved. Our neighbors had bought their sets to watch the wedding of Crown Prince (now Emperor) Akihito (b. 1933) and Shoda Michiko (b. 1934) in April 1959. By 1964, nearly half of the roughly 25 million households in Japan owned a television, the wealthier among them a color model.
Once we had a TV, I began to watch two cartoons that starred robots: Tetsuwan Atomu (Astro Boy) and Tetsujin 28 [Nijuhachi]-go (Ironman 28, aka Gigantor) (figure 1). Both were preceded by comic book versions and broadcast on Fuji TV from 1963 to 1966. Gigantor is remotely controlled by a ten-year-old boy detective whose father created the robot. The boy and the robot share a deep emotional bond that underscores the familial aspects of real-world human–robot relations in Japan, a theme I will reiterate in each chapter. I discuss Astro Boy at length in chapters 4 and 5, but it is worth noting in this connection that in addition to human friends, the boy-bot has a robot family of his own.
When interviewing Japanese roboticists, our shared childhood fascination with robot cartoons helped to break the ice and spark illuminating digressions. I also noticed that virtually all of them had either a picture or a figurine of Astro Boy somewhere in their office or lab. Later, I learned that many of their students and younger colleagues, while familiar with Astro Boy — arguably Japan's most famous cartoon robot — watched Doraemon during their childhood. Doraemon is a blue and white, bipedal robotic cat with a huge smile that lives with a human family and is especially close to the preteen son (see chapter 5).
These cartoon robots, like most of their counterparts in science fiction and the hundreds of "robot films" produced since Metropolis (1927) by Fritz Lang (1890–1976), are dexterous and possess superhuman powers. In comparison, actual, tangible robots seem clumsy, slow, and underwhelming. For example, the state-of-the-art robots participating in the June 2015 DARPA Robotics Challenge "accomplished tasks at a glacial pace" and often toppled over. Like the video of that event, footage of actual robots moving are typically speeded up significantly, sometimes ten to thirty times their original speed. They are also heavily edited to create the illusion of smooth, coordinated movement.
The industrial robots employed in the automotive industry in Japan, Europe, and the United States are efficient and precise, and supercomputers equipped with AI (artificial intelligence), like IBM's Watson (1997) and Google DeepMind's AlphaGo (2016), can beat human chess and Go players, respectively. Futurists such as Vernor Vinge (b. 1944) and Ray Kurzweil (b. 1948) speculate that robots with the sensorimotor versatility of the average human will fully evolve between 2029 and 2045 — an outcome they refer to as "the Singularity," the moment in the future when humans and machines will converge. Paul Allen, cofounder of Microsoft, is more guarded and argues that the "amazing intricacy of human cognition should serve as a caution to those who claim the Singularity is close." "Heady stuff," he declares, "but a very long time coming" (Allen and Greaves 2011).
Some skeptics dismiss the possibility of the Singularity altogether. For example, Erik Sofge (2014a) contends that it should be recognized for what it is: "a secular, S[cience] F[iction]-based belief system." Although a Singularity Institute in Japan was established online in 2014, its five founding members are not well known, if known at all, among Japanese roboticists (Japan Singularity Institute 2014). Kurzweil, in contrast, is a member of (and has been honored by) the Institute of Electrical and Electronics Engineers (IEEE), the international organization to which most active roboticists belong. Singularity, as a state of being or form of existence, does not seem to have much of a profile or following in Japan. Generally speaking, what Japanese roboticists and their public- and private-sector supporters are interested in pursuing is not the convergence of humans and machines, but rather the coexistence of humans and robots.
WHAT IS A ROBOT?
In their 1952 "critical review" of the concepts and definitions of culture, anthropologists Alfred Kroeber (1876–1960) and Clyde Kluckhohn (1905–60) collected 156 examples. Today anthropologists tend either to consider the term culture self-evident — we always already know what we mean by it — and/or to provide new variants for any one of those 156 examples (Kroeber and Kluckhohn 1952). And so it is with the term robot.
Masahiro Mori (b. 1927), the Japanese roboticist who first introduced the concept of the "uncanny valley" (which I will discuss and demystify in chapter 6), claimed in an interview that
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. [Honda's] 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. (Mori, cited in Kageki 2012)
Similarly, Illah Nourbakhsh, a professor of robotics and director of the CREATE Lab at Carnegie Mellon University, writes in Robot Futures that one should "never ask a roboticist what a robot is. The answer changes too quickly. By the time researchers finish their most recent debate on what is and what isn't a robot, the frontier moves on as whole new interaction technologies are born" (Nourbakhsh 2013: xiv).
With these caveats in mind, I will provide a working definition of robot by first considering its etymology. The English word robot derives from the Czech robota (drudge laborer). Coined by litterateur Karel Capek (1890–1938) and his artist brother Josef Capek (1887–1945), the word first appeared in the former's play, R.U.R.(Rossumovi Univerzální Roboti [Rossum's Universal Robots]), published in 1920. A science fiction melodrama with comical passages, R.U.R. is about a factory (Rossum's Universal Robots) in the near future where artificial humans are mass produced, from protoplasmic batter, as tireless workers for export all over the world. To make a long story short, new-model robots, provided with emotions and able to experience anger at their perceived exploitation, revolt en masse. They kill all but one human, a traditional artisan. Since the formula for the batter has been destroyed, robots cannot reproduce themselves in the factory. Instead, the artisan encourages an emotionally enhanced robot couple he calls "Adam and Eve" to go and repopulate the world with their own kind.
The R.U.R. robots are indistinguishable from flesh-and-blood humans except that they lack ethnic differences. This fact, the director of Rossum's realizes too late, means that unlike humans, for whom ethnic and national differences provide a pretext for war, the monoethnic R.U.R. robots are able to unite en masse to slaughter all humans (Capek [1921] 2004: 46). Two distinct categories of robot bodies are assembled at Rossum's factory: female and male. The factory's director explains why female models are needed, in effect distinguishing sex from gender. He notes that the factory is simply responding to customer demand for robots that conform to gendered occupations: female robots are needed as "waitresses, shop-girls, secretaries" (22). Capek's robots thus reinforce the self-evident (and binary) construction of the sexual and gendered division of labor in human society. Not only do robots today look different from one another, and even have nationality, most are also gendered.
R.U.R. was performed in Tokyo in 1924 under the title Jinzð Ningen (Artificial Human). The play, along with Lang's Metropolis, which was screened three years later, sparked an ongoing fascination in Japan with robots in popular culture. This interest continued apace with cartoonist Osamu Tezuka's Tetsuwan Atomu (Astro Boy) in the 1950s and is evinced today by the humanoids, animaloids, and cyborgs that dominate manga (cartoons) and anime (animated films). From the 1920s to the present day in Japan, robots have been cast as both threatening and helpful to humans, but mostly the latter. Even before the 1960s, when the state embarked on a policy of automation in lieu of replacement migration to extend the productivity of the domestic workforce, the general trend in Japanese popular media and culture has been to characterize robots as benign and human-friendly. Capek's graphic portrayal in R.U.R. of the end of bourgeois humanity at the hands of a violent robot-proletariat helped to shape the fears of "westerners" about the destructive potential of robots that persist to this day. The dystopian play did not, however, compromise the mostly favorable acceptance among Japanese of things mechanical, including robots, from the 1920s onward. Today, whether benevolent or malevolent, the meaning of "robot" has become closely associated in Japan and elsewhere with intelligent machines whose shapes and functions are inspired by biology and diverse life-forms.
As already noted, roboticists resist defining what exactly a robot is. Of all the many definitions available, I find the following one usefully comprehensive yet concise: A robot is an aggregation of different technologies — sensors, lenses, software, telecommunication tools, actuators, batteries, synthetic materials and fabrics — that make it capable of interacting with its environment, with some human supervision (through teleoperation) or autonomously. Regarding this last criterion, robots are not (yet) completely autonomous. Rather, autonomy exists on a sliding scale: the level of autonomy is adjusted according to different scenarios and influences the way that humans and robots interact with one another (Beer et al. 2014: 74).
Irrespective of its level of autonomy, to be called a humanoid, a robot must meet two criteria. One, its body must resemble a human being, having something like a head, arms, torso, and legs. Two, it has to perform in a human-like manner in environments designed for the capabilities of the human body, such as an office, hospital, school, or house. As I discuss at length in chapter 4, most Japanese humanoids are gendered. In humans and robots alike, gender — femininity and masculinity — constitutes a repertory of learned and performed behaviors and gestures that are cosmetically and sartorially enhanced. Some humanoids are so lifelike that they can actually pass as human beings. These gendered robots are called androids (male) and gynoids (female).
At the turn of the twenty-first century, Japanese engineers were the first to prioritize the development of humanoid robots. The bones of our human ancestors were discovered in Tanzania's Olduvai Gorge, and Japan is a cyber–Olduvai Gorge, where humanoids first emerged and where they continue to evolve. It should be clear from the array of robots featured in this book that, unlike the generic humanoids of R.U.R., robot morphology is just as diverse as that of humans. They come in every size, shape, and color. All of the robots referred to in this book are enormously complex, layered systems and represent an amalgamation of research across and within many disciplines, from electrical engineering to child development studies.
Some roboticists regard humanoids as nostalgic throwbacks. General-purpose humanoid robots, they say, are too complicated and expensive to use in industry and should be more practically replaced by modular robots with specific functions — which, like Lego blocks, can be joined and separated as needed for a task (Devenish 2001). Even roboticists at Japan's National Institute of Advanced Industrial Science and Technology (AIST) acknowledge that "it is not easy to develop the next-generation robot industry, especially that of biped humanoid robots. The major barriers for industrialization include: (1) robots walking on two feet only have little commercial value, (2) the unit price is very high, and (3) if it falls, it may be seriously damaged" (AIST 2009).
Similarly, Kenji Hara, a senior analyst at the Japanese research and marketing firm Seed Planning, remarks that many of Japan's robotics projects tend to be too far-fetched, concentrating on humanoids and other "leaps of the imagination" that cannot be readily brought to market (Tabuchi 2009). These critiques are in a minority in Japan, in the sense that they are not widely publicized in the mass and social media, which instead sensationalize the latest gee-whiz humanoid robot. Honda's ASIMO (figure 2), Mitsubishi's Wakamaru (figure 3), Hitachi's EMIEW, and SoftBank's Pepper (figure 4) have all enjoyed their fifteen minutes of fame. Public relations videos crafted by robotics manufacturers are broadcast on television as if they were documentaries of actual, real-world human–robot interactions. Even the IEEE, the world's largest and most prestigious technical professional society, includes on its website a weekly selection of "awesome robot videos." The marketing hype around the nearly four-foot-tall, sixty-one-pound Pepper and other humanoids effectively creates expectations that cannot be met, especially among consumers whose image of robots has been forged by science fiction films, anime, and comic books.
Pepper was introduced in 2014, with much media fanfare, by Masayoshi Son (b. 1957), CEO of SoftBank, as a "personal emotional robot" companion. The following year, SoftBank's website announced that "since the 20th of June 2015, Pepper has become the first humanoid robot available to Japanese households. Pepper is their new daily companion!" (SoftBank Robotics 2016a). About ten thousand Peppers have been purchased by robot hobbyists and some businesses (e.g., Pizza Hut and Nestlé) seeking a novelty attraction. Retailing for an average of $1,700 each, Peppers are sold on "a subscription contract that includes a network data plan and equipment insurance," totaling $360 per month, which drives up the cost of ownership. Robots are expensive machines. Pepper is equipped with a pricey array of technology, including multiple cameras, microphones, and depth sensors that let Pepper make eye contact and respond to touch. Due to the robot's comparatively rudimentary software, however, it responds only to a narrow range of specific preprogrammed cues, rather than to the highly variable facial and vocal expressions made by humans. SoftBank is willing to sell each Pepper at a loss — at the price of a pet dog (Alpeyev and Amano 2016; Singh 2015) — in anticipation of an expanded market for robot companions in a few years. But will there be such a market?
As Frank Tobe wrote in October 2016, Pepper has "failed in every way to (1) be a companion, (2) recognize emotional cues, (3) be able to converse reliably and intelligently, and (4) provide any level of service other than first time entertainment." His bleak assessment corroborates my own underwhelming experience interacting with Pepper on several occasions at SoftBank's flagship store in Harajuku (Tokyo). Depending on the question, the boy-bot either launched into a seamless pitch (in Japanese) for SoftBank products or froze. By August 2016, the Harajuku store had limited customer interaction with Pepper to swiping the iPad on its chest. Several robot engineers left SoftBank after they realized that the robotics division was headed by project managers with little knowledge of robotics or AI. SoftBank public relations and media hype need to be (re)evaluated against these sobering realities (Tobe 2016).
(Continues…)
Excerpted from "Robo sapiens japanicus"
by .
Copyright © 2018 The Regents of the University of California.
Excerpted by permission of UNIVERSITY OF CALIFORNIA PRESS.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.
