Note: All materials on this site are the copyrighted property of Alfred B. Bortz. Individuals may print single copies of reviews or columns for their own use. For permission to publish or print multiple copies of any of the materials on this site, please contact the author by e-mail.
What is intelligence, and can people build intelligent machines?
Those are not new questions. People began asking them seriously at least as early as the eighteenth century when skilled technologists built automata that mimicked human and animal actions. Three new books explore those questions from very different perspectives.
FROM AUTOMATONS ...
Tom Standage's story,
The Turk: The Life and Times of the Famous Eighteenth-Century Chess-Playing Machine (Walker and Company, 288 pages, $24.00),
begins in 1769 when Austro-Hungarian Empress Maria Theresa summoned thirty-five year-old Wolfgang von Kempelen, known for his technological skills, to observe a performance of a visiting French conjuror and his automatons.
Throughout the show, Kempelen chatted with the empress, explaining the workings of the contrivances. At the end of the show, Kempelen declared himself "capable of constructing a machine, the effect of which would be much more surprising, and the deception far more complete, than anything the empress had just witnessed."
Six months later, Kempelen presented an automaton in a distinctive costume that could not only pick up and move pieces on a chessboard but also play the game expertly. "The Turk," as the creation became known, was soon a sensation. On and off, it traveled the world for eighty-five years, generating publicity wherever it went and a stream of articles claiming to have deduced its secret.
Mr. Standage's chronology holds the attention of twenty-first century readers, even though to them it is obvious that the automaton required a human operator. The book reveals the Turk's secrets and goes beyond to the 1997 triumph of IBM's Deep Blue over world Chess Champion Garry Kasparov.
... TO ROBOTS ...
The Turk closes with speculation about the future of thinking machines. The author disagrees with robot experts Raymond Kurzweil (The Age of Spiritual Machines) and Hans Moravec (the unforgettably disconcerting Mind Children and Robot) about their "essentially numerical arguments about the inevitability of machine intelligence."
MIT Professor Rodney A. Brooks, who as a teenager built his first robot almost exactly two centuries after the Turk's debut, also disputes Moravec's conclusion that robots will be our evolutionary successors, but for different reasons.
In Flesh and Machines (Pantheon Books, 288 pages, $26.00), he describes Moravec as "a true eccentric. Brilliant, innovative, and nuts. He was a tremendous influence on my life...."
That statement alerts readers to challenges ahead. They lean forward, eager to challenge Dr. Brooks at every step. Chapter by chapter, his well-organized stories and well-crafted language prod them relentlessly towards a brash prediction of, as the subtitle states, How Robots Will Change Us. Chapter by chapter, they reconsider their own position, never quite accepting the author's view but never quite able to dismiss it, even as it veers toward the radical.
In the end, after having seen artificially intelligent behavior emerge from a colony of insect-like robots; after having seen people respond socially to machines that display human-like emotion -- even when they know the underlying technology in detail; after wondering what makes humans special or if robots can be special in the same way; after all of that, readers follow Dr. Brooks out on a limb.
They watch him leap into a future not of intelligent robots, but of human-computer hybrids, when robotic implants will give ordinary people extraordinary sensory capabilities -- even Internet-mediated telepathy and telekinesis.
Like spectators marveling at the Turk's skill, readers look at their robotic destiny and almost -- almost -- believe it will all be possible.
A new approach to exploring that question may come from Synaptic Self: How Our Brains Become Who We Are by Joseph LeDoux (Viking, 400 pages, $25.95). Although Dr. LeDoux makes no reference to artificial intelligence, preferring to focus on real brains in great detail, his book may be valuable to computer engineers who study the properties of circuits known as neural networks.
Those engineers view the brain as a programmable network of interconnected neurons whose interconnections can change as the result of experience. Dr. LeDoux's detailed presentation may provide them with a number of useful models. He argues that hereditary and environmental influences determine the strength and complexity of the neural connections at synapses. A human self is thus the product of genetic endowment and life experience as reflected in the synapses of that individual's brain; it is the manifestation of chemically mediated electric flows through those connections.
VIVE LA DIFFERENCE
If humans and robots merge, will any distinctions remain? Most of us hope so. Those few who expect to be reincarnated as machines would do well to consider the following Cyborg's Lament.
I am thinking; therefore I exist.
But there's one part of life I have missed.
Could it be that my skin
Made of plastic and tin
Is the reason I've never been kissed?