by David Levy
Love and Sex with Robots
The Evolution of Human-Robot Relationships
David Levy
To “Anthony,”* an MIT student who tried having girlfriends but found that he preferred relationships with computers. And to all the other “Anthonys” past, present, and future, of both sexes.
According to the United Nations Economic Commission for Europe’s World Robotics Survey, in 2002 the number of domestic and service robots more than tripled, nearly outstripping their industrial counterparts. By the end of 2003, there were more than 600,000 robot vacuum cleaners and lawn mowers, a figure predicted to rise to more than 4 million by the end of next year. Japanese industrial firms are racing to build humanoid robots to act as domestic helpers for the elderly, and South Korea has set a goal that 100 percent of households should have domestic robots by 2020. “Probably the area of robotics that is likely to prove most controversial is the development of robotic sex toys,” says Dr. Christensen. “People are going to be having sex with robots in the next five years,” he says. “Initially these robots will be pretty basic, but that is unlikely to put people off,” he says. “People are willing to have sex with inflatable dolls, so initially anything that moves will be an improvement.”
—The Economist, June 8, 2006, quoting Henrik Christensen, chairman of the European Robotics Network at the Swedish Royal Institute of Technology in Stockholm
Contents
Epigraph
Introduction
Part One: Love with Robots
1 Falling in Love (with People)
2 Loving Our Pets
3 Emotional Relationships with Electronic Objects
4 Falling in Love with Virtual People (Humanoid Robots)
Part Two: Sex with Robots
Introduction to Part Two
5 Why We Enjoy Sex
6 Why People Pay for Sex
7 Sex Technologies
8 The Mental Leap to Sex with Robots
Conclusion
Notes
Searchable Terms
Acknowledgments
About the Author
Other Books by David Levy
Credits
Copyright
About the Publisher
Introduction
Recent research shows that people perceive and treat robots not just as machines, but also as their companions or artificial partners.
—Alexander Libin and Elena Libin, 20041
At the dawn of the twenty-first century, mankind is experiencing an era of phenomenal scientific and technological achievement. Whole disciplines of science that were unheard of even a few decades ago are now making possible amazing feats in areas such as cell-phone technology, computer technology, space research, and medicine. Furthermore, our scientific knowledge is growing at a rate that is itself increasing. The more we know about a science, the more quickly we may use our knowledge to discover even more within that science. This has been very much the case in the field of computing, a science that (like me) was in its infancy in the early 1950s. In those days each of the few computers that had been built would fill a room and cost a fortune. And although articles about computers appeared from time to time in the popular press, few people had any idea what these newfangled machines could be used for. When, in 1943, an American company called International Business Machines first considered the possibility of manufacturing computers on a commercial basis, the company’s founder and president, Thomas J. Watson, pessimistically predicted, “I think there is a world market for maybe five computers.” How wrong he was! Instead of the computer’s being something of a commercial white elephant, it became the product for which IBM is best known. And by 1981 the computer had become so ubiquitous in industry, in the office, and in academic life that IBM launched a whole new product category called the personal computer, the PC, a computer that was not only more powerful than the multimillion-dollar machines of twenty years earlier but was also affordable for many families and individuals.
Commensurately with this dramatic growth in the popularity of the computer as a tool for all to use, computer science became a subject that was increasingly studied at universities and research institutes. And within computer science there came an even newer discipline, called artificial intelligence,* the science of making computers that can think. Every science has its own divisions and subdivisions, and artificial intelligence (AI) is no exception. Developing programs to play games such as chess falls within the boundaries of a division of AI called “heuristic programming.”† Programs that carry on conversations or translate from one language into another are encompassed within the AI discipline of “natural language processing.” And among the other disciplines within AI there is robotics.
The word “robot” was suggested by Josef Capek‡ in discussion with his more famous brother, the Czechoslovak writer Karel Capek. It is derived from the Czech robota (forced labor) and was first revealed in the West when Karel used it in the title of his play Rossum’s Universal Robots (R.U.R.), an immediate hit when it was first shown on Broadway. The literal meaning of “robot” is “worker.” The robots in C? apek’s play were creature machines, resembling humans in appearance, designed and built to serve as workers for their human masters.
Although the word “robot” was new in the early 1920s, the idea of an artificial form of life was by no means a new one in Capek’s day. Inventors and engineers had for millennia devised automata that simulated some of the functions of living creatures. One of the earliest to do so was Heron§ of Alexandria, who lived in the first century A.D.** Among many inventions that were mechanical marvels for their time, Heron constructed some water-powered mechanical birds, entire flocks of them, that even emitted realistic chirping sounds created by a water-driven device.
The public’s fascination for automata reached its first peak in France in the eighteenth century. One example of this genre was a menacing mechanical owl set amid a group of smaller birds, designed in 1644 by the French engineer Isaac de Caus. The smaller birds would flutter their wings and chirp while the owl slowly moved on a pivot to face them. As the owl’s face turned toward the smaller birds, appearing to threaten them, they became still and stopped their chirping. When the owl’s face then turned away from the group, the smaller birds came alive again. The whole mechanism was driven by a water wheel that controlled the actions of each bird by means of a metal cylinder, the surface of which was embedded with pins, just like a music box. As the cylinder turned via the force of the water, the pins on the cylinder would engage with a music box–like mechanism so that each pin created its own effect or movement in one of the birds.
Following de Caus’s example, at least two other French automaton inventors also used birds as the embodiments for some of their mechanical marvels. In 1733 an inventor named Maillard designed a mechanical swan that would paddle through the water while its head moved slowly from side to side. Maillard’s idea was as simple as it was clever: a paddle wheel, similar to those found in the Mississippi River steamboats, propelled the swan forward while simultaneously connecting, via a system of gears, with the swan’s head; as the paddle wheel rotated, it thus served a dual purpose, creating the forward motion of the swan’s body and the simultaneous side-to-side motion of its head. An even more advanced idea, and a more entertaining example of this genre, was a mechanical defecating duck, the creation of Jacques Vaucanson. The duck could bend its neck, move its wings and its feet, and it could “eat.” It would stretch out its neck to peck at corn offered by a human hand, then swallow, digest, and finally excrete it, the corn having
been turned into excrement by a chemical process, according to Vaucanson. In fact the “digestion” and “excretion” processes were parts of a hoax. The corn, once eaten, was held in a receptacle at the lower end of the duck’s throat, while the duck’s “excrement” was not genuine duck droppings but some other material that had been inserted in the duck’s rear end prior to the demonstration.
That Vaucanson’s duck did not actually digest its food and defecate in no way diminishes its contribution as a precursor to humanlike robotics. One of the principal achievements of Vaucanson and his peers was the stimulation of widespread interest in the mechanical aspects of what is now known as artificial life. That era saw the creation of automata that could not only eat but also breathe; automata with soft skin, flexible lips, and delicately moving jointed fingers.* A remarkable example of a humanoid automaton was a birthing machine designed in the mid-eighteenth century by Angélique du Coudray, midwife to the royal court of France. The purpose of this machine was to assist in the teaching of midwifery, as a result of which many examples of the machine were made and sent to doctors and midwives throughout France. Du Coudray’s machine was made of wicker, stuffed linen and leather, dyed in various flesh-tone colors, some pale and some of a deeper red, to simulate the softness and appearance of a woman’s skin and organs. The pelvic bones of human skeletons were used in some of her machines, and sponges soaked in liquids colored red and other hues were used inside the machine, releasing their simulated bodily fluids at appropriate stages of the lectures on the birthing process.†
While Vaucanson and his peers managed the simulation of physiological and other natural bodily processes, there were other inventors who focused on simulating the processes of thought. One of the best known of these peers was, like Vaucanson, also famous for a machine that turned out to be a hoax. In the closing years of the eighteenth century, Baron Wolfgang von Kempelen, a scientific adviser to the royal court of Vienna, designed a chess-playing automaton in the guise of a Turk seated on a wooden box. Despite Kempelen’s assurances to the contrary, and his magician-like demonstrations to convince his audiences that the wooden box contained nothing untoward, there was in fact a (small) strong human player secreted in the box, a player who vanquished all chess enthusiasts who tried their luck against “the Turk.”
On the far side of the world, the Japanese interest in robotics also dates back to the eighteenth century, during the Edo period in Japanese history, with the design of a tea-carrying doll, called karakuri. When a host, seated opposite his guest, placed a cup of tea in the doll’s hands, it carried the cup to the guest, who then took the cup from the doll, whereupon the doll stopped moving. After drinking the tea, the guest put the cup back into the doll’s hands, the weight of the cup causing the doll to spin around and return to the host with the empty cup. These dolls were fashioned in the form of a child, with the technology hidden inside, creating an aura of mystery and magic. Rather than designing their automata to look like animals, as many of the French inventors had done, the Japanese had realized more than two hundred years ago that automata are more appealing if presented in the guise of humans, a realization that anticipated some of the research described here in chapter 4.
These eighteenth-century marvels did much to create a climate of interest in the notion that human and animal bodily and mental processes can be successfully simulated. By 1830, walking dolls were being constructed and exhibited in Paris, and soon thereafter came dolls with moving eyes. Next came dolls that could eat, drink, dance, breathe, and swim (in three different strokes: backstroke, breast stroke, and crawl). And for those that could drink, one inventor, Leon Bru, created an artificial bladder, so that after taking a drink his dolls could pee. It was in this climate, and with the benefit of the nineteenth-century development of electricity, that the idea of robots as we now see them began to take root.
Karel Capek’s vision was of robots that could think for themselves, robots with feelings, robots that could fall in love with each other. In Rossum’s Universal Robots, one of the scientists at the robot factory came up with the idea of endowing the robots with emotions, which led to their developing feelings of resentment about being treated like the slaves of human masters. Capek had the foresight to predict what some people today fear about a future with robots—that they will “take over the world”—and in his play, the robots decided to rebel and kill all human beings.
When it premiered in New York in 1922, Rossum’s Universal Robots was hailed by one critic as a “brilliant satire on our mechanized society,” and the concept of robots as Capek envisioned them was taken up by several science-fiction writers, most notably Isaac Asimov. In 1940, Asimov reacted to the plethora of books and stories that had already been published in which man created robots that became killers. Asimov proposed three “laws of robotics,” later augmented by a fourth law, all designed to safeguard mankind’s interests in the face of whatever ideas the robots of the future might develop.*
Since the birth of the science of artificial intelligence in the mid-1950s, gigantic strides have been made in the quest for a truly intelligent artificial entity. The defeat of the world’s best chess player, Garry Kasparov, was just one of these strides. Others include the creation of computer programs that can compose music that sounds like Mozart or Chopin or Scott Joplin, at the operator’s behest; programs that can draw and paint better than many human artists whose work today hangs in art galleries and in the homes of wealthy collectors; and programs that can trawl the Internet and write news stories based on the information they gather, stories written in a style of which most journalists would be proud. Then there are expert systems—programs that incorporate human expertise to enable them to solve analytical problems normally assigned to human experts. Such programs are powerful tools for medical diagnosis, and they have also proved to be highly competent in a wide diversity of other fields, such as prospecting for minerals, making political judgments, detecting fraudulent uses of credit cards, and making recommendations in court cases to judges and lawyers, even advising defendants how to plead. These are not examples of what might be in the future—they are just some of the accomplishments of AI in its first fifty years. During the second half of the twentieth century, science fiction became a hugely popular literary form, paralleling the development of the science of artificial intelligence. One exemplar of this parallel is the computer Hal in Arthur C. Clarke’s 2001: A Space Odyssey. Hal crushes David, the human hero, at chess, mirroring the defeat of Garry Kasparov, four years prior to 2001, by IBM’s Deep Blue chess-playing computer.
It was industry that prompted the initial Japanese research into robotics. And although it was also in industry that robots were first employed to replace humans ( just think of car factories), the major thrust of robotics in Japan during the 1990s and into the first few years of the present century has been in “service” robots. At first, service robots were mainly used for drudgery-related tasks—cleaning robots, sewer robots, demolition robots, mail-cart robots, and robots for a host of other tasks, such as firefighting, refueling cars at gas stations, and in agriculture. But after the service-robot industry became well established in Japan, the country’s robot scientists turned their attentions to the realm of personal robots, to be used at home by the individual. Mowing the lawn and vacuuming the carpet have both become tasks that in a slowly but steadily increasing number of homes are now undertaken by robots. Similarly, robots are beginning to be used in education, and Toyota has announced that by 2010 the company plans to start selling robots that can help to look after the elderly and to serve tea to guests in the home. This trend, from the use of robots in industry to their use in service tasks and now in the home, represents a shift toward an increasing level of interaction between robots and humans. In industry a button is pressed and the robot springs into action on the assembly line, working away on a repetitive task with little or no need for supervision until the daily quota of cars or whatever has been manufactured. If a robot can manage as
sembly-line tasks once, it can manage them time and again. If your car works well when you buy it, you can reasonably assume that the next guy’s car will also work well, and the next, and so on. That is the great advantage of industrial robots—not only do they do the job as often as is needed, they do it as well the hundredth time, and the thousandth, as they did the first time. And it is this advantage of repetitive excellence that makes the industrial robot so impersonal.
A service robot does not normally need to perform its designated task time and again, one immediately after another. Instead it is there like a butler, to be at the beck and call of the individual when needed to mow the lawn or vacuum the floor, a task that might occur only once a day, once a week, or even less often. But to use a service robot requires of its owner much more interaction than with industrial robots. The owner often needs to collaborate with the robot—by bringing it onto the lawn, for example—before the robot can start work, and then to wheel the robot away again when its task has been completed. Not always, however. Some lawn-mower robots take themselves off to the garden shed when it rains or when their work is done, and some even recharge themselves by wandering over to the power socket and connecting themselves when their batteries are low—an electronic parallel of “I’m hungry, Mommy, so I’m going to take some food from the fridge.”
As with many other lines of research in robotics, the first fully working androids (human-shaped robots) were developed in Japan. Development on androids started at Waseda University in the 1970s, many years before the states of the art in computing, vision technology, and various branches within artificial intelligence reached the levels needed in a twenty-first-century autonomous android. The 1980s saw a burst of engineering effort in artificial hands and other limbs, but at the time there were very few industrial applications for such technologies, and so the momentum from those efforts was not sustained throughout the 1990s. But after a gap of a decade or so, Waseda University and other Japanese robotics groups are now making good use of that earlier research-and-development effort.
