These are my “lecture notes” for the lecture that I gave at the Art University in Linz, Austria (at the Interface Cultures Invited Lectures Series) on June 21, 2011.
An Interdisciplinary Approach to Building Robots
by Alan N. Shapiro
In the first part of my lecture, I presented the idea of beginning a multidisciplinary and “humanities informatics” approach to the development of robots-slash-androids, going beyond the mechanical engineering approach to ASIMO’s advancement that Honda has taken so far.
(The lecture notes for the first part of my lecture are published at my website as “Towards a Unified Existential Science of Humans and Androids”.)
I think that Honda has made ingenious innovations so far in the Honda robot project, which recently celebrated its 10th anniversary, in 2010. They have succeeded in making a robot which “operates in the human living space” and is people-friendly. That is an excellent starting point. HONDA doesn’t want us to be afraid of robots. ASIMO is especially good at walking and running. It can leap, and absorb the impact of landing. The “intelligent, real-time, flexible-walking” technology known as i-WALK enables ASIMO to “walk continuously while changing directions” and to have “great stability in response to sudden movements.” ASIMO can interpret the body language of humans, greet people, follow people, and walk in a direction that humans indicate. ASIMO has face, sound, and voice recognition capabilities. It can call a person by name, shake hands, or walk hand-in-hand with them. It has built-in visual, ground, and ultrasonic sensors. It can carry a tray or push a cart.
In some of their literature, Honda is now saying that the next step in ASIMO’s development is going to be AI (so-called Artificial Intelligence). But Artificial Intelligence, as it has been pursued in the academic field of Computer Science and by advanced labs in computer industry companies, has already failed for the last 60 years. In my view, Artificial Intelligence is not going to happen until there is a revolution in the foundations of Computer Science. Together with my partner, the Austrian software architect Bernhard Angerer, I am working on conceptualizing and implementing this revolutionary transition from Computer Science 1.0 to Computer Science 2.0. The main basic principle of the upgrade is to stop regarding software objects as inert things to be handled in an industrial process, and start to view software objects as living entities. Once we reinvent Computer Science starting from the insight that software objects are alive, then we will indeed achieve Artificial Intelligence, the igniting of the autonomous capabilities of software objects. To continue to regard software objects as non-living is simply a big mistake, in my opinion. We are making this same big mistake in a lot of areas. In our system of food production, we treat animals as non-living thing-objects to be handled in an industrial process. This should be changed. And we don’t want to view robots in this way either, as not having subjectivity.
I said that it will be important, thinking from the perspective of Science Fiction Studies, to understand the negative associations with robots that exist in the mind of the public, and to elaborate a marketing strategy that instead connects with a comprehensive set of alternative positive associations about androids. In this second part of my talk, I will outline a programme of 7 concrete near-term steps that a large robotics manufacturing company should take on the road towards introducing robots-slash-androids to the consumer public.
First, we should study the physical and choreographed movements of dancers in order to gain understanding of the highly flexible body-in-motion as the basis for the prototype artificial humanoid.
Here I would like to draw attention to the work of the Finnish philosopher of technology Jaana Parviainen on the phenomenology of dance in a real/virtual environment. In her essay, “Dance and Digital/Virtual Technologies”, which I published at my website, Parviainen does a phenomenological analysis of the learning process of a group of dancers who worked with the EGM (Embodied Generative Music) interface, a software created at the Institut für Elektronische Musik und Akustik in Graz, Austria. “Motion capture and sound processing software is combined with movement improvisation and performance,” Parviainen explains. Precisely what Honda should do to bring forward the future physically agile development of the ASIMO robot: have the robot learn how to move from the hybrid software and dance medium. Exactly what the dancer learns, and exactly what the robot wants to learn as well, is bodily knowledge, knowledge of the body.
To take advantage of this golden opportunity, Honda would have to be open to seeing the value of an academic knowledge that comes from the humanities, not from engineering and what is conventionally in English called science. Parviainen applies the phenomenology of perception of Maurice Merleau-Ponty and the phenomenology of “dwelling” of Martin Heidegger to explore how the dancers “explore reversibility between sound and movement to shape this [ceaseless interchange] in an artistically expressive manner.” Knowledge of the body and its movements is a largely unexplored territory, and dancers are the true pioneers of this consciousness. I hope that Honda will become aware of this reality and have the robots follow the dancers. Parviainen explicates the details of how bodily knowledge is built up in a dance practice within a virtually-enhanced environment at the crossroads between the motoric and the sensory dimensions. She considers the pleasures of movement in kinaesthesia (muscle sense), tactility (sense of touch), proprioception (awareness of the position of the body), and body memory.
Second, we will want to have a software development project for improving the dialogic-discursive dimension of human-machine speech interaction technology.
The speech recognition component of Spoken Dialogue Technology is already well-developed in its functionality, since its refinement derives from engineering-intensive acoustic modeling algorithms. For the language domain where parsing of verbal utterances into discrete word identifiers is desired, the software builds a stochastic or “Hidden Markov” statistical model speech patterns reference database. But the language understanding component of Spoken Dialogue Technology, which has the functional requirement of assigning a meaningful representation to grammatically correct or colloquially loosely-structured sequences of words, and passing a coded value onto the dialogue manager subsystem, is still relatively primitive in its growth. It will be much more difficult to work up language understanding into a viable technology that can support the conducting of a conversation that is at a qualitatively rich enough level that it resembles a natural-language interaction between two humans. The syntactic and semantic analyses involved in interpreting the sense and non-sense of the declaration are a painstaking process.
Additionally, we have not yet even begun to introduce knowledge of the poetic elements of language into the analysis of meaning at all. We have not yet taken seriously as an information theory the study of the subtle differences in signification – both in the nearly-silent whisper of the signifier and in the content-meaning of the signified – among the iterated-enumerated elements of cross-language chains of associated words and sounds, as elaborated in the science (of radicalized semiotics) known as Grammatology that was first systematized by the philosopher Jacques Derrida. Derrida was an information theorist without a computer. Applied grammatology will be the correct, productive, and fruitful way to deal with the ambiguities of language which exist on the lexical plane of grammar, in the intricacies of sense, and in the real world pragmatics of structurally decentered and chaotic communication.
Third, we must undertake the very subtle work of the preliminary design (and, later, programming) of the emotion chip (like the one that the android Data of Star Trek: The Next Generation gets in the episode “Brothers”).
As the Italian historian of American literature and culture Franco LaPolla says, “what characterizes Star Trek: The Next Generation is the fact that we feel and experience more deeply and more often the problem of identity of the android Data, and not the situation of any other living being on the starship. On more than one occasion – for example, in the very beautiful episode ‘The Measure of a Man’ – the show expresses clearly the allusion from Data to Pinocchio.”
In the amphitheater-like courtroom, Commander Riker begins the trial to determine if Data has human rights by making a brilliant case, based largely on Data’s own deposition, that his dear friend is nothing but a machine. Sitting in the witness chair with his hand over the identification scanner, NFN NMI (no first name, no middle initial) Data testifies that he has a maximum storage capacity of 800 quadrillion bits, and a total linear computational speed of 60 trillion operations per second. He can effortlessly bend a plasteel rod packing a tensile strength of 40 kilobars with his bare hands. Riker dispassionately removes Data’s left hand and forearm from his body to show their internal electro-mechanical composition. With this action, Riker symbolically robs Data of the appearance of human subjectivity. “Its responses are dictated by an elaborate software written by a man, its hardware built by a man, and now a man will shut it off,” the prosecutor proclaims.
Riker flips a switch in Data’s back, just below his right shoulder blade, and the android collapses into unconsciousness. “Pinocchio is broken, its strings have been cut.”
In the episode “Data’s Day,” the android asserts: “If being human is not simply a matter of being born flesh and blood, if it is instead a way of thinking, acting and feeling, then I am hopeful that one day I will discover my own humanity. Until then, Commander Maddox, I will continue learning, changing, growing and trying to become more than what I am.”
The uncertainty regarding the ultimate humanity of Data is a historical-existential necessity for us human viewers at the dawn of the 21st century. Paradoxically, Data’s condition mirrors our own radical uncertainty today in nearly losing yet trying to regain the ability to truly place ourselves in the psychological condition of feeling oneself as human. What is so interesting about the android condition of Lt. Commander Data is that it discloses the extreme difficulty of the human condition as it really is today under hyper-corporate-capitalism. More than merely the search for humanity, it is the challenge which all of us now face in an obligatory confrontation with ourselves that we can no longer evade.
In “The Measure of a Man,” where the question of Data’s relationship to humanity is on trial, it is the sage Guinan, played by Whoopi Goldberg, who provides the solution to a dejected and demoralized Captain Picard, Data’s defense attorney, while he is taking a short break in the Ten-Forward lounge.
Sitting together late at night in the deserted recreation room, Guinan hints to the Captain that the real issue of the trial is slavery. The hearing’s true significance is the imminent danger of long-term subjugation by the United Federation of Planets of a race of expendable creatures who would do society’s “dirty work” and menial tasks. If the arrogant Maddox and his kind have their way, the black-skinned El-Aurian sage suggests, the harrowing outcome will be “an army of Datas, all disposable. You don’t have to think about their welfare. You don’t have to think about how they feel — whole generations of disposable people.”
Captain Picard suddenly recognizes that Guinan is talking about the rebirth of slavery. If he loses, the decision made at this hearing will establish the precedent of all future Soong-class androids being regarded as nothing but property. It is not just about Commander Maddox being granted authorization to carry out his disassembly procedure. It is about the fate of all the future Datas that Starfleet will build should Maddox or some other robotics scientist succeed. It is about the act of humanity degrading itself by treating its humanoid technological creation in such an instrumental way. Slavery, says Picard, is “not a word we want back in our vocabulary.”
Picard returns to the courtroom and his place next to Data. Inspired by Guinan’s insight, he magnificently turns around the basic issues of the trial. He opens up searching questions about the nature of the Federation and ourselves. What would declaring Data to be property say about us? What kind of beings would we be if we define androids in this condescending manner? How will we be “judged as a species” if we behave towards our creation in this way? “If they’re expendable, disposable, aren’t we?” Picard makes it clear that what we think about Data will “reveal the kind of a people we are.”
The key ideas for devising the parameters of the android emotion chip are to be found in the practice of Gestalt Therapy as developed by Fritz Perls, Laura Perls, Paul Goodman, Jerry Kogan, and Wiltrud Krauss-Kogan, especially in the cultivation of existential knowing that is enabled through answering the question “how do you feel?” in the present moment.
Compared to classical Freudian psychoanalysis, or Jungian analytical psychology, or the Lacanian school, Gestalt Therapy has received relatively little attention among leftist intellectuals in the Western countries, and in the academic fields of critical sociology and cultural studies. This is a huge mistake. And it is especially surprising, given that one of the founders of Gestalt Therapy was Paul Goodman, who was a major American anarchist social theorist. However, the paradox is that it is very difficult, and even perhaps undesirable, to try to systematize or codify the ideas of Gestalt Therapy into any kind of academic writing. In his autobiography called In and Out The Garbage Pail, Fritz Perls talks about his seminal ideas in a performative and personal way, even using lengthy poems and jokes as modes of writing. I think that this book and the life-experience that it recounts are very important for the android consciousness and the awakening of emotions. To be completely and freely in the world one must freely express one’s feelings and emotions. In the existing repressive order of the dominant society, that is hardly possibly, but we must find ways. One of Fritz Perls’ key ideas is to replace order with Gestalt. Gestalt organizes life, but in a less straightjacketed way. Perls felt that both Freudian and Reichian theories downplayed the emotions. “Nature is not so wasteful as to create emotions as a nuisance. Without emotions we are dead, bored, uninvolved machines.” Gestalt Therapy also emphasizes the importance of self-supporting breathing and independence. Breathing is essential for the life of the organism. The experience of Gestalt Therapy is about contact, about involvement and engagement with life and with the other (I and Thou). Contact stimulates a greater appreciation of differences.
Fourth, breakdowns, accidents, injuries, and minor illnesses must be programmed into the life of robots, in order to increase their resemblances to the well-functioning imperfect health system that is the human being, and therefore aid in the expansion of a new medicine.
One of the cultural theorists whom I am most interested in is the French thinker Paul Virilio, whose work on technology has focused on architecture, art, transportation, war, urban planning, and the cinema. Virilio is also a theorist of accidents and crashes, and we need to carefully consider his ideas about accidents in undertaking the design and building of robots. According to Paul Virilio, every technology has both a rational, utilitarian purpose and a built-in accident “waiting to happen.” The accident is intrinsic or requisite, not incidental, to the given technology. But we can undo this inevitability of disaster by reversing the engineering goals of the system with respect to the evaluation of perfection and imperfection. By consciously making systems and artefacts which are less-than-perfect, we can in fact make safer technologies. The human body and the maintenance of human health are just like this, in my opinion. It is good to occasionally have things go wrong and to have minor ailments. It is good to have a fever once a year to exercise the immune system. It is good to have a cold in order to cleanse the system, and to protect the body against more debilitating conditions like viruses, or things which would require hospitalization. Androids and humans will learn how to be physically and mentally, medically and psychologically, imperfect together.
Fifth, there should be a mobility project that researches the relationship of human and android bodies to the dynamic physical systems in which they dwell, such as trains and planes, and shopping malls and department stores.
In the area of architecture, urban planning, physical environments, and mobility, I believe that we are living through a fundamental transition from the era of human beings considered as things to be handled and managed in an industrial process to an era of awareness that the human being is truly alive and has an embodied dwelling in space-time and an existential being-in-the world. The stressful and alienating conditions of modern life have made it a categorical ethical imperative that we create architectural spaces and transportation systems with the more sensitive ontological subjectivity of human beings in mind. Planes, trains, and automobiles, for example, should be redesigned for the experience of the riders rather than merely as vehicles for getting from point A to point B. Shopping malls and department stores are a major part of our cultural and economic life, all over the world. Yet they are not achieving the cultural or economic goals which they could achieve, since they are not yet designed from a broad interdisciplinary perspective. The experience of foreign and international cultures should be part of shopping malls, and education should be integrated into the experience of shopping. The way that we interact with nature, and with virtual realities and simulations should be seen as important dimensions of the architecting of these ambiences of commerce and everyday life.
Sixth, there should be consideration of android (poly)sexuality.
Prior to the 1970s, we lived with a system of sexuality where heterosexuality was the norm, and other styles of sexuality, like homosexuality, were regarded as deviant. In the 1970s, following the sexual liberation movements of the 1960s, a new system of sexuality was instituted in Western culture. The system of recognized sexual identities was established. It was decided that 5 kinds of sexual identity are acceptable: heterosexual, gay, lesbian, bisexual, and transsexual. This system of the 1970s was a major improvement for human freedom and happiness as compared with the previous system. But this system is now 40 years old and it is obsolete. We need new ideas about sexuality. Polysexuality wants to build a bridge between the heterosexual majority and homosexual minorities. Polysexuality is about the sexual liberation of the majority, learning from practices and subcultural styles of queer minorities.
The main difference between the view of sexuality in existing heterosexual and homosexual ideologies, and in Polysexuality, is that heterosexual and homosexual ideologies regard sexuality as being an identity. Polysexuality regards sexuality as a desire. Polysexuality regards sexuality as fun. It is nothing more than fun and desire and experiences. Sexuality is a performance, as Judith Butler says. It should be a series of temporary performances. To regard sexuality as an identity is existentially inauthentic. It is to practice “bad faith,” exactly as the great mid-20th century existentialist philosophers Jean-Paul Sartre and Simone de Beauvoir defined that important term. Simone de Beauvoir, the author of The Second Sex, was probably the most important feminist thinker of all time. “Bad faith” is when I escape from the authenticity of the difficult open-ended questions of the human condition by taking refuge in the security of an identity which is, in fact, a simplification and a reduction, the false imposition of an abstract idealist idea onto materialist-existentialist reality.
Seventh, the technology of the audio-visual capture and storage of memories will be deliberated, with an eye towards later on giving memory implants to our partner species.
How would it be possible to achieve memory implants for robots-slash-androids, similar to how that technology is depicted in the science fiction film Blade Runner (based on a novel by Philip K. Dick)? And how would it be possible to achieve memory implants for humans, similar to how that technology is portrayed in the science fiction films Total Recall (based on a short story by Philip K. Dick) and Johnny Mnemonic (based on a short story by William Gibson), where the title character, played by Keanu Reeves, has a cybernetic information-storage memory chip brain implant? Clearly, in order to do these two things, we will need a new Computer Science, which I call Computer Science 2.0, that is based on the first principle that software is alive, as opposed to Computer Science 1.0, which is based on the unaware unspoken assumption that software is an inert machine entity. The difference between the two Computer Sciences comes down to one famous line from Philip K. Dick’s novel Ubik: “I’m alive and you’re dead.”
In order to manufacture an artificial equivalent of human memories, we will need to go way beyond the current techniques of audio-visual digital recording. In order to develop the technology of memory implants, we will need to get inside the human brain and understand how it works. Computer Science 2.0 is bio-informatics and informational biology. It is at the crossroads between computing and biology, something that will surpass both of these fields of knowledge as they are presently constituted. Software for us at Shapiro Technologies is both something man-made and something natural. There already exists a software that is alive, and this is the human brain. Here we are in the realm of something squishy – it has sometimes been called wetware rather than software. We will actually require samples of it in order to replicate it. We will need miniaturized nanotech-bots to get inside the brain, sort of like in the science fiction film Fantastic Voyage. We we also need to comprehend it subjectively through consciousness. We will need to go on a psychedelic trip like the astronaut played by Keir Dullea at the end of the science fiction film 2001: A Space Odyssey. We will need to read Marcel Proust’s Remembrance of Things Past. And we are already designing the nuts and bolts of Computer Science 2.0 as the logical successor to the earlier paradigms of procedural-functional programming and object-oriented programming. To grasp the living informational system of the brain, the digital-binary principles of Computer Science 1.0 simply won’t do.
Now I want to say something about my favorite movie of all time. It is called Blade Runner.
What makes Blade Runner extraordinary is that it artfully presents an alternative to the two predominant ways in which artificially intelligent machines or androids are thought about and depicted in mainstream techno-culture. These modes recur again and again in novels, scientific pundit books, and Hollywood films. For theorist-entrepreneurs like Ray Kurzweil (The Age of Spiritual Machines) or movies like A.I. Artificial Intelligence (2001), Bicentennial Man (1999), or The Matrix (1999), there are two possible ways of imagining Artificial Intelligence. Either it is a question of androids attaining human-like characteristics (computational skills, memory capacity, emotions, intuitions, behavior, and consciousness), and therefore accepting to have as their goal to become equivalent to humans. Or it is about androids exceeding human intelligence and skillfulness, and therefore becoming an ominous menace to humanity as they seek to dominate us. Never is it about humans and androids co-existing in difference or, better, otherness, alterité, Andersheit.
For mainstream techno-scientific thinking, it can never be a question of peaceful co-existence in otherness because there can be only one master of the universe. The story of life and biological-technological evolution, for someone like Ray Kurzweil, is a “billion-year drama that led inexorably to its greatest creation: human intelligence.” It is an economic thinking of the “Darwinian” (what a misreading of the great interdisciplinary thinker and literary writer Charles Darwin!) competitive battle of the “survival of the fittest,” applied universally and extended indefinitely into the past and future. It is the achievement of a so-called intelligence that enables the Western technological domination of nature and other species (animals).
The job of Blade Runner Rick Deckard, played by Harrison Ford, is to weed out, hunt down, and retire trespassing replicants who have surreptitiously made their way back to decaying Earth society from their slave labor assignments in the off-world colonies or on space exploration expeditions. Deckard is a technical expert at distinguishing android skin jobs made by biotech companies like the Tyrell Corp. from human beings. But the resonating message of the film of ideas Blade Runner is that we are all replicants.
The uncertainty of Deckard’s ontological status as human or replicant is brought out more forcefully in Blade Runner: Director’s Cut (1992), which restores an uncanny twelve-second dream sequence of a majestic silver-white unicorn running through misty woods, shown when Deckard nods off while playing the piano. Lt. Gaff, who makes origami figures, leaves the tiny tinfoil form of a unicorn on the floor just outside Deckard’s apartment in the film’s final moments. The juxtaposition of dreamland and decorative variants of the mythical equine creature delicately hints that Gaff and the police authorities know the content of Deckard’s dreams. The divorced sushi lover’s dreams and wishes have been technologically implanted, just as he himself knows of Rachael’s childhood recollection of the baby spiders outside her window, which was the technical reproduction of a memory of Tyrell’s niece.
I conclude with a quote from the late Franco La Polla, from the chapter “Data and Baudrillard,” from one of the three books of his great trilogy of Star Trek analysis:
“Il robotico, per tornare a dove siamo partiti, non va letto tanto come una ricerca di perfezione, ma piuttosto come una nostalgia di essa (il Roy di Blade Runner essendone probabilmente l’immagine più alta e intensa), con l’aggiunta di una certezza: che, anche se attinta, essa non coinciderà mai più con quella originaria (di qui la connessione con la minaccia e il pericolo a volte proposta dalla sua immagine, dalla sua figura). Data, l’androide perfetto del cervello positronico, incarna proprio la consapevolezza di questo: l’umanità, nelle sue evidenti contraddizioni, s’identifica nel grado ultimo di perfezione cui egli aspira. Data è una delle maschere dell’immaginario contemporaneo, il vero umano di tutto il quadro proprio in virtù della sua ricerca di umanità, della sua identità ogni volta soggetta a uno scarto, a una inadeguatezza, a una domanda.”
“Robotics, to return to where we started, is not to be read as a search for perfection, but rather as a nostalgia for it (Roy Batty of Blade Runner being probably the highest and most intense image of this). Even if this perfection is attained, it will never coincide again with the original (and from this stems the connection with the threat and the danger that at times its image and its figure represent). Data, the perfect android of the positronic brain, incarnates precisely the awareness of this impossibility. Humanity, in its obvious contradictions, identifies itself with the ultimate stage of perfection to which it aspires. Data is one of the masks of the contemporary imagination, the true human of THE BIG PICTURE precisely due to his search for humanity, to his identity that is constantly subjected to rejection, to an inadequacy, to a question.”
