Robots that "care" are no longer merely science fiction ...
Producing machines that look and behave like people seems to be a human project with a long history. Mention of a Jewish Rabbi producing an instance of the legendary golem (a creature understood to possess an active human-like body, while lacking a soul) appeared as early as the 4th century CE. The celebrated 13th-century Muslim engineer Ibn al-Razzaz al-Jazari undoubtedly designed, and may have constructed, what has been described in present-day terms as "the first programmable human-like robotic device" – a spectacular artifact featuring four robotic musicians performing on a floating boat (Nicks 2010). Inspired by animated life-like figures reportedly created by an ancient Greek named Ctesibus, Leonardo da Vinci – around the time in the 1400s at which he began painting his famous Last Supper – also designed a human-like robot resembling a knight in armor. Fascination with the idea of crafting convincing imitations of people appears to have been part of human history for millennia.
In more recent times, though, modern computers – and with them, research introducing so-called "artificial intelligence" (AI) – have begun to give this long-standing fascination some significant new dimensions. Perhaps the most widely recognized contemporary human-like (or, nowadays, "humanoid") robot is a product of Japanese science and technology named "ASIMO." Resembling a short (4 ft 3 in) astronaut wearing a backpack, ASIMO represents the fruit of several decades of research and development conducted by the Honda Motor Company. Videos on the company's official web site show ASIMO climbing stairs, jogging, balancing on one foot, visually recognizing people by name, and serving a tray of beverages to restaurant patrons. Similar examples of this impressive humanoid robot technology exist in other countries as well – e.g., Turkey (Today's Zaman 2010), United Arab Emirates (Fahad Inc. 2008), and South Korea (Impactlab.net 2008).
Investment by business enterprises in the significant cost and engineering effort required to design and build these curiously humanoid machines constitutes one of the "new dimensions" previously mentioned. Historical figures such as Al-Jazari and da Vinci, after all, were not responding to global marketing prospects with their robotic creations. In contrast, a current Honda Motor Company web site tells us that ASIMO was intended to be more than an attention-catching novelty from the beginning; in fact, it was "created solely to perform tasks to assist people, especially those lacking full mobility" (Honda Robotics 2011). Similarly, the president of a Robotic Industries Association reports that South Korea is "taking the lead in promoting the use of robots for service applications such as elder care" (Burnstein 2009). A former GM of the Microsoft Robotics Group has identified such assistive care as the market that "intrigues" him the most, citing approaching increases in senior populations – and, consequently, heavier burdens upon healthcare systems – as factors that may "present the 'killer app' for personal robots" (Foley 2009). A 2009 online report titled "Robot Nurses to Care for Japanese Elderly within Five Years" reports that Warwick University, in England, has undertaken a "three year 2.7 million dollar project to develop a robot nurse," predicting that "nurses could be delegating tasks to robotic colleagues by 2020" (Zygbotics 2009).
It is important to note that such robotic "colleagues" of human nurses commonly are intended to be suited for fairly intimate kinds of social interactions with people. One finds, for instance, references to robotic assistance in recreation and with feeding, grooming, walking, bathing, etc. (Babyboomercaretaker.com 2007). Accordingly, we encounter another new dimension. Robotic arms have welded and painted in our automobile factories for decades, but the repetitive activities of these familiar industrial robots are profoundly different from interaction with a humanoid machine that helps one's aging grandmother eat her dinner and take her medicine (perhaps even chatting and playing a card game with her). Moreover, the latter type of robot no longer is mere science fiction; design and construction of machines to perform these kinds of personal human-robot interactions are taking place now.
Considered only as machines meant to assist overburdened nurses with their care of older people, the types of humanoid robots just described might initially be categorized simply as useful new tools. We have reasons to wonder, though, how long those who will be interacting regularly with these life-like robots can be expected to perceive them merely as tools. So-called "animaloid" robots, such as the robotic dog AIBO that was marketed in recent years by the Sony Corporation, admittedly represent a somewhat different class of robotic artifact than the more complex contemporary humanoids such as ASIMO. Nevertheless, empirical studies of human-robot interaction even with AIBO have uncovered some relevant thought-provoking surprises. Not long ago, for example, numerous online postings by owners of AIBO began appearing on Internet forums. One study of these postings noted the following confession by an AIBO owner:
The other day I proved to myself that I do indeed treat him as if he were alive, because I was getting changed to go out, and [AIBO] was in the room, but before I got changed I stuck him in a corner so he didn't see me! (Friedman, Kahn, and Hagman 2003, 278)
Regardless of whether this posted confession was altogether truthful, its expressed thought of needing modesty in this case clearly alerts us to the potential psychological potency of human interaction with such machines. Abrahamic religions, through their shared accounts of the Garden of Eden, have long recognized appropriateness of modesty between even the primordial man and woman – but application of that sentiment to our dealings with a battery-operated dog suggests how plastic human notions of personhood might be!
For that matter, professional testimony of such plasticity for the specific case of humanoid robots is available in a frequently-quoted set of observations by Professor Sherry Turkle, Director of the MIT Initiative on Technology and Self, at the Massachusetts Institute of Technology. One of her MIT colleagues, widely-recognized roboticist Rodney Brooks, is among the many people who have cited Turkle's report of her first encounter with his experimental humanoid robot, Cog; note carefully Sherry's candid description of the experience:
Cog "noticed" me soon after I entered its room. Its head turned to follow me and I was embarrassed to note that this made me happy. I found myself competing with another visitor for its attention. At one point, I felt sure that Cog's eyes had "caught" my own. My visit left me shaken – not by anything that Cog was able to accomplish but by my own reaction to "him." For years whenever I had heard Rodney Brooks speak about his robotic "creatures," I had always been careful to mentally put quotation marks around the word. But now, with Cog, I had found the quotation marks had disappeared. Despite myself and despite my continuing skepticism about this research project, I had behaved as though in the presence of another being. (Brooks 2003, 149)
Professor Turkle's testimony is consistent with an entire literature of contemporary research in human-robot interaction that suggests a deep human predisposition progressively to accept as peers various machines that convincingly mimic human appearance and autonomous behavior. Her reference to discovering herself behaving as though she were "in the presence of another being" points, in turn, toward some questions that invite our reflection.
First, one might inquire whether (and why) it could matter that humans seem so inclined to regard convincingly humanoid machines as peers. For some people, it apparently does not matter. From his perspective as a practicing Zen Buddhist, for example, robotics engineer Masahiro Mori has argued against insisting upon any profound distinction between persons and robots, noting that there "must also be buddha-nature in the machines and robots that my colleagues and I make" (Mori 1999, 174). In contrast, though, a pilot study has suggested that Abrahamic theistic belief in creation of individual human souls by a personal deity may be related to disapproval of human-robot interaction "with life-like personal robots that requires human acceptance of the robots at intimate levels" (Metzler and Lewis 2008, 22). This finding resonates with a respected voice in modern Christian theology. Paul Tillich, in Volume Three of his monumental Systematic Theology, addresses "objects that are produced by the technical act," warning that "by virtue of producing and directing mere things" one can lose one's "character as an independent self" and "become a thing" (74). Again, Jewish theologian and philosopher Martin Buber, widely remembered for his distinction between "I – Thou" and "I – It" relations, issues a similar warning in I and Thou:
And in all the seriousness of truth, hear this: without It man cannot live.
But he who lives with It alone is not a man. (34)
Apparently, we have reasons to expect that individuals belonging to Abrahamic religious traditions may especially feel troubled when they find themselves treating humanoid machines as though they were peers.
Within the Abrahamic religious family, after all, human beings historically have been regarded as spiritually special, and understood as belonging to a category fundamentally different from any technological artifacts that they might construct for amusement, or as tools. Anglican priest (and physicist) John Polkinghorne has emphasized significance of "the mystery of the human person," which involves "our embodied nature, embedded in the physical world but transcending a merely reductive physicality" (Polkinghorne 1998, 80). Both the mystery and the transcendence that Polkinghorne mentions are punctuated clearly, as well, in the Holy Qur'an: And they will ask thee of the Spirit. SAY: The Spirit proceedeth at my Lord's command: but of knowledge, only a little to you is given (The Night Journey – Sura 17:85). The theistic perspective of this family of religions tends to link the human person, as a free moral agent, with a spiritual level of reality that is not completely expressible in terms of everyday (macro-level) entities such as rocks and trees – and machines.
It may be pertinent at this point to inquire whether the spiritual level of reality envisioned by these religious faiths might arguably be represented even in current science. To be sure, the robotic and AI technologies upon which we have focused in this essay are discussed almost entirely nowadays with so-called "macro-level" accounts of discrete, individualized entities. Computer scientists typically view all "information processing" executed by contemporary computers as reducible to operations of the celebrated Turing Machine formalism, which imagines an abstract machine successively "reading" well-defined symbols (0 or 1) on an external tape, comparing them with its current internal "state," and then implementing clearly prescribed (albeit possibly null) changes on the tape and its own internal state. Physicists working with quantum mechanics, however, have discovered a quite different level of reality that requires a so-called "quantum-level" description. The description is expressed mathematically in terms of complex numbers (incorporating an imaginary unit equal to the square root of negative one) and it explores a reality in which individualized entities of the macro-level (this table, that book, etc.) simply are no longer present. An atom may be understood to contain four electrons, but – in principle – one cannot select and track, say, the changing locations over time of a specific individual electron among the four. Pondering this strange new reality, mathematical physicist Roger Penrose has argued (via his Shadows of the Mind) that human consciousness cannot be modeled in terms of the Turing Machine formalism, requiring, instead, the resources of an advanced quantum physics. If the emerging technology of "quantum computers" eventually could yield a machine consistent with Roger Penrose's understanding of how the human brain operates when we experience consciousness, future robots incorporating such computers might open possibilities for exciting new dialogue between religion and science.
Under present circumstances, though, we can discern the outlines of potential difficulties in the not-so-distant future. Specifically, elderly members of the Abrahamic faiths may find themselves increasingly conflicted in responding to robotic "caregivers." On one hand, following natural predispositions, they will be inclined to accept the machines as caregivers (dropping the skeptical quotation marks, as Professor Turkle did during her encounter with Cog). At the same time, they may retain their religious worldviews and resist accepting the machines as persons. Will they feel authentically comforted, then, by machines programmed to display "artificial empathy"? Will they discover resolution of their conflict in the following conjecture by noted roboticist Hans Moravec?
So, it may be appropriate to say "God" has granted a soul to a machine when the machine is accepted as a real person by a wide human community. (Moravec 1999, 77)
Indeed, in perhaps the next ten years or so, how will people be using quotation marks to distinguish what they consider authentic from mere "make-believe"? Will they be describing new robot nurses as persons – or as "persons"? Will they decide that the machines care for people – or "care" for people? Will the artifacts be considered capable of moral behavior – or "moral" behavior? Will some older people still understand the granting of souls to be determined by God – or by "God"?
For some of us, these already are important questions.
Theodore Albert Metzler is the Director of Darrell W. Hughes Program for Religion and Science Dialogue, Oklahoma City University.