Information processing will never spawn human conciousness in computers (good anti-AI rant from Jaron Lanier)greenspun.com : LUSENET : Human-Machine Assimilation : One Thread
Information processing will never spawn human conciousness in computers.
COMPUTERS ARE IMPROVING FASTER than any technology people have experienced before, so the obvious question is: Where is all this leading? It is not enough for us, as some thinkers are doing, to throw up our hands and announce that computers are already out of control. That's just a way of avoiding responsibility.
Lately, we've been asking some momentous questions about computers. Are they on the way to becoming a new life-form? Are they a successor species to people? Will people merge with them into something entirely new? What is unsettling is that, given how fast computers are improving, some of these predictions might be tested while many of us are still alive.
What sort of thing would consciousness have to be to distinguish the present moment?
Jaron Lanier Right now, two fast-changing technologies are already challenging what it means to be a person.
Biotechnology offers the potential for reproductive techniques and genetic manipulations that will force us to redraw the circle around what is human. Information technology does the same thing through artificial intelligence. The center of the circle that defines a person is a dot called consciousness, and as murky as that subject is, we are fast approaching some crucial conclusions about it. So far, one idea has received more avid support than any other. This is the notion that computers are becoming more and more "alive" and capable of judgment. It has become a clichi of technology reporting and a standby of computer industry public relations. It is a myth that depends upon support from, of all people, intellectuals.
A small but robust cadre of intellectual shock artists these days is paid to assert that people aren't real. Every year one sees more books, mislabeled as popular science, aimed at furthering the claim that the reader is an illusion: The Astonishing Hypothesis; Consciousness Explained; The User Illusion; How the Mind Works. On and on, the formula repeats itself. An arrogant, overreaching title is essential, as is an overview of recent brain research to create an atmosphere of scientific authority. But the science is peripheral to the boast that the author has attained a superior vantage point from which to decimate the reader's tender sense of self.
At first it might appear that these authors are talking about people, not computers, but what they really are doing is talking about all of reality as if it were only information processing. That, in turn, puts the superlatively fast computer of the not-too-distant future in a godlike position.
Mainstream computer science is comfortable with the notion that computers are becoming "intelligent." The origin of this idea can be found in the celebrated thought experiment known as the Turing Test, named after Alan Turing, one of the inventors of the modern computer. According to Turing, if a computer and a human are concealed behind screens, and a human judge who interacts with both is fooled into thinking the machine is a human, then the computer can be said to possess an intellect equivalent to a human being's. At the heart of this test is the concept that if an entity is made of chips instead of biological cells but can act the same way as a person, consciousness may not really matter. Many thoughtful computer scientists, influenced by Turing, believe this is so. If the Turing Test is right, it would seem unfair, even "racist" in a way, to think of a computer as a mere tool for humans.
But in practice, the Turing Test is dangerously flawed because it is impossible to distinguish whether the computer is getting more humanlike or the human is getting more computerlike. People are vulnerable, unfortunately, to making themselves stupid in order to make computers appear to be smarter. In the real world, the Turing Test is mundanely played out in miniature every day. For example, many people will change their financial behavior in subtle ways to appear favorably to credit card and bank computer programseven on occasion borrowing when they don't need to in order to make the right "impression" on their bank. The Turing Test has a comparable, if different, distorting effect on intellectuals. It helps anticonsciousness authors claim that personal experiences do not define the center of a person. Thinkers throughout history, from Zen poets to existentialists, have suggested that an individual's everyday experience is an illusion, but the anticonsciousness authors go further: They want to convince the reader that experience itself is an illusion.
Our inability to demonstrate our inner experience is what makes the consciousness-free point of view so compelling. It is wicked fun, apparently, to upset soft, fuzzy, consciousness-bearing people because they can't easily respond. Moreover, if we doubt the consciousness of a humanized (Turing Test-passing) computer, we can, and must, face the terror of doubting the legitimacy of every consciousness we presume to contact, except our own. We are back with Descartes in the steam bath. Leaving behind God and souls, the new test of faith becomes the existence of other people, even in everyday life. (The existence of one's own consciousness remains the sole thing that need not rest on faithbut what a lonely, solipsistic position to be in.)
But we don't have to play the game of the anticonsciousness crowd. Instead, we can ask questions that take us down a different line of reasoning. First: What would the universe be like if there were no consciousness? One answer, which I do not accept, is that the universe would instantly vanish if consciousness were removed. Another is that it would be about the same, except people would act a little more rigidly. A third possibility is that the universe would remain in an undetermined state without consciousness to cause quantum "collapses."
I propose the following answer: Without consciousness, the universe would be exactly the same, in terms of what its particles would be doing. A camera placed in a consciousness-free universe would record exactly the same scene. The same photons would travel the same courses through the lens. What would be different is that there would be no basis for recognizing gross objects. The particles of a piece of wood would be in just the same locations at the same moment, but the notion that the collection of particles is also a piece of wood would not be put forward. There would be no minds, words, or meaning in such a universethey are extraneous human interpretations. And that suggests a question that anticonsciousness smart alecks should be made to answer: Where do objects come from?
Of all the objects we would be unable to refer to if consciousness did not exist, the most compelling, I believe, is the concept of time the idea of the present moment. That's the second question: Where does the now come from? Even if it were possible to subscribe to the fuzzy notion that consciousness "emerges" from certain kinds of adjacencies of particles, the present moment remains clearly dependent on the spooky, dualistic kind of subjective consciousness we all sense is really there.
In fact, consciousness is a sensible proposition only in the context of time. Certainly consciousness would be a pointless exercise if the next moment were never a surprise. But the existence of the present moment has been taken for granted. It's true that a lot of attention has been given to its imperfect nature. From Edmund Husserl's explication to contemporary cognitive science experiments and relativistic theory, the present moment has been shown to be a tattered, faulty entity. But it is still there. And the fact that we perceive it at all demands an accounting.
Instead of focusing on the simplicity of the present moment, philosophical inquiries into the relationship between consciousness and time have tended to center on free will. Indeed, it might at first seem that free will and consciousness must sink or swim together. Both are embraced by "folk wisdom" but have been forced to withstand sustained assault by a great many mainstream scientists and philosophers in the 20th century. If one of these twin concepts is shown to be an illusion, the other might be expected to be equally vulnerable. But it turns out that just the opposite is true: If free will is taken to be an illusion, then consciousness must fill a daunting and heroic epistemological role.
If the universe is rigorously deterministic, and free will debunked, the only feature left to define the present moment is consciousness. If it were not for conscious experience, there would be no way to distinguish the present moment from other moments. Unless, that is, the present moment can be shown to be "doing something" distinguishable on its own, which would mean that there is something nondeterministic about the way the universe moves from one moment to the next. The universe itself would have to gain something very much like free will to fill the role played by consciousness in identifying the present moment. Either way, something is going on beyond the boundaries of empiricism.
It is easy to demonstrate that consciousness is localized in space, specifically in portions of the brain. But consciousness is even more strikingly placed in time. We seem to experience the present, remember the past, and anticipate the future. This is extremely important. Neither the brain nor any other thing in the universe possesses an objective trait by which one can distinguish the present moment. Right here...now...is discernible from other moments only by someone who is stuck in time by being conscious.
What sort of thing would consciousness have to be to distinguish the present moment? In a deterministic universe, it would be something quite different from mere synapses shooting sparks in the brain, for those also are simply events in the deterministic flow. Thus they cannot mark the present moment as being different and separate from some other moment. Instead, a strategy of dualism must be invoked to account for the ability of anyone to speak of the present moment in a deterministic universe.
Sure, the present moment is not a precise point but a messy approximation, fraught with compromise and illusion. And indeed, people have a rather poor sense of the absolute passage of time. The neural events concerned with how we perceive the present moment are distributed in different parts of the brain, creating delays in the way our minds coordinate the information needed to tell time with total exactitude. But even if the present moment could be known with precision by a single mind, there would be no way in a relativistic universe for other minds to single out the same moment. Relativity does not allow for an absolute clockyet we are all sharing the moment.
We are getting into complicated material here. The critical point is, however, that while consciousness cannot distinguish the present moment with consistency or precision, there is no other candidate in a deterministic universe that can distinguish it at all. No matter how accurate a timepiece is, including a computer's internal clock, it can't tell the present moment from another moment. That task falls to the conscious, responsible observerto us. And that is what makes us conscious and human.
Jaron Lanier is a computer scientist, composer, visual artist, and author. He coined the term "virtual reality" and was a pioneer in scientific, engineering, and commercial aspects of the field.
-- scott (email@example.com), February 16, 2000