Computers and Human Identity

Transcription

1 Computers and Human Identity As computers work their way ever more deeply into our lives and work, they will become part of our identities. Who we say we are and the way we think will be influenced by our computers. We will become the technology we have created. As profound as this observation may be, it is an old human trait. Scientific models of the mind mimic the dominant technology of their time. Rene Descartes's mind-body dualism, for instance, proposed a spirit inhabiting a mechanical device resembling an extraordinarily complicated clock-the "ghost in the machine?' Freud's theories of the unconscious implicitly evoked a steam engine: impulses blocked from their natural release would build up pressure in the subconscious mind until they blew out elsewhere, far from their point of origin. In the last half century, our models of the mind have become electrical and computational. Today, neuroscientists routinely talk about "feedback loops" and "brain circuitry?' Instinctive behaviors are said to be "programmed" or "wired"; we possess the "software" for certain kinds of mental activity. These usages are not only metaphorical, they are part of

2 88 Beyond Calculation: The Next Fifo/ Years of Computing everyday practice and conversation. It is entirely respectable for a cognitive scientist in the 1990s to maintain that all cognitive processing, including processing of visual images, takes place via abstract, language-like propositions, which sound a great deal like algorithms. Unlike earlier technological models of the mind, the algorithmic one is complicated by the reality that we have spent much of the past fifty years trying to make computers that resemble the mind. As the steam engine was an iron horse, the computer has been envisioned as a silicon human. Computer scientists have gone beyond replicating our ability to calculate; they have attacked object recognition, language comprehension, reasoning about the world, playing chess, and much more. Given this history, it is remarkable that the authors in this section either ignore or disavow the idea that computers can or should mimic human abilities. In so doing, they have given voice to a modern paradox: brains are computers, but computers are not brains. This paradox suggests a complex symbiosis. It is as if a fungus managed to invent its own algae: the resulting lichen has new attributes and extends into new dimensions that neither fungus nor alga would have achieved on its own. Similar to us, yet not quite mirrors, computers are our newfound partners in creating something never before seen on earth. Here are some glimpses of what the first stages of this new, amalgamated metaorganism will be like, and who we will be, cohabiting with them. In "Growing Up in the Culture of Simulation;' Sherry Turkle, a computer media expert, sociologist, and anthropologist, describes the profound and surprising ways in which the children to whom she has listened so carefully think about life-nonlife distinctions. Her eight-, nine-, and ten-year-old subjects, largely innocent of the notions of the precomputer age, show us how the presence of evolving "life-forms" in games like SimLife are training them to be comfortable with fluid, mutually distinct, and sometimes even contradictory ideas of the fundamental properties of life. They are of an age at which one can still learn to speak a foreign language without an accent; an age at which talent may be fostered into virtuosity. These children will ultimately manipulate computers, and the ideas underlying them, with a facility and apparent effortless-

3 Computers and Human Identiry 89 ness that will leave their elders gaping with awe. If we want to know how people will think about computing in the future, Turkle's subjects are well worth our attention. Donald Norman, a psychologist who has studied computers and people for many years, compares the thought styles of the two symbiotic partners, human and computer-one unpredictable, robust, relatively error-insensitive, and redundant; the other abstract, linear, consistent, rational, and precise. The increasing complexity of modern life stems in large part from our attempts to accommodate ourselves to a thought style for which we never evolved and which we are not especially good at. Anyone who has had to interact with a mainframe computer, master a command-based operating system like DOS, keep track of a bewildering number of passwords, or even program a VCR knows that it is not an intuitive process. Rather than continually try to adapt to a world in which we feel awkward and ill at ease, we can look for ways to make that world accommodate itself to us. We can leave to computers the things they are good at-calculation, memory, precise and flawless repetition-and reserve for biological humans the pattern-based, language-based, sensory tasks they do best. In counterpoint to Norman's discussion of why it's good that computers don't work like the brain, David Gelernter might have titled his essay "Why It's Bad That Computers Don't Work Like the Brain?' In "The Logic of Dreams:' Gelernter, a professor of computer science at Yale, proposes a radical idea that is in its way one of the more conservative ideas in this book. He is still pursuing the goal of a machine made in the mind's image. Not satisfied with distinctions such as Norman has drawn between the natural capacities of computers and humans, and unwilling to throw up his hands and admit defeat in modeling any mental capacity, Gelernter looks at one of the most distinctive of all human attributes, the ability to glean new ideas out of undirected, seemingly random associations-in other words, creativity-and suggests how such a capacity could be modeled in a computer program. Franz Ait is a founder and past president of the ACM and a pioneer of the computing field. In "End-Running Human Intelligence:' he returns us to the dominant strategy for thinking about computing capabilities: how we can

4 90 Beyond Calculation: The Next Fifty Years of Computing take advantage of what computers do well. As if in reply to Gelernter and in support of Norman, Alt writes, ''The danger lies in any attempt to let AI imitate the same strategies as are followed by the human mind in solving a problem." He compares such efforts to nineteenth-century attempts to invent a machine that would reproduce human handwriting. Since writing by hand was obviously more efficient than setting lead type, it was dearly the more promising model to follow in trying to produce machine writing. But the machine that ultimately succeeded, the typewriter, employed a mechanical strategy that in no way resembled handwriting. Alt suggests a number of tasks, many of them surprising, where entirely nonhuman strategies could outperform human intelligence: the drawing up of contracts; medical diagnosis; weather forecasting; and conducting public opinion surveys. From the distribution of abilities comes the distribution oflabor. Will computers one day take over every conceivable task, as envisioned by Jack Williamson's 1954 science-fiction story "With Folded Hands"? Paul Abrahams, a programming language expert, software engineer, private consultant, and former ACM president, takes a skeptical look at this suggestion in ''The World Without Work?' Comparing the prospect with an attempt to build a tower to the moon ("Each year [the builders] point with pride at how much higher the tower is than it was the previous year"), he also asks the very useful question of how far we want to go in this direction. Finally, in "The Design of Interaction;' Terry Winograd examines three current trends in the human-computer symbiosis and speculates on where they're taking us. "Computation to communication" means not only the increasing connectivity of the now worldwide network but also the shift in usage from calculation to communication. "Machinery to habitat" means that what people are most aware of when they interact with a computer has changed over the years from the machine as a piece of hardware, to the operating system, to an application program, to most recendy a fully realized virtual world. ''Aliens to agents" refers again to this section's dominant theme: the transformation of our vision of computers from incipient beings like ourselves, if only we could manage to give them birth, to helpers with which together we will create an astonishing new world. Winograd predicts that the

5 Growing Up in the Culture of Simulation 91 new approach to design called for is more likely to be taken up by a new field, interaction design, than by computer science. He suspects that the identity of a "soft designer" does not fit with the inclinations of computer scientists to be hard technologists.

6 Figure 7.1. All the World's a Stage.lUchard Rosenblum, Iris inkjet print.