Last Sunday saw the conclusion of the historic Chess match between Garry Kasparov and IBM's Deep Blue computer, in which for the first time the world chess champion lost to a machine under normal competition game rules. Deep Blue eventually lost the match, but not without honor. The media are hailing this event as a turning point in human-machine relations, but it diminishes in significance if we consider how much we have yet to achieve in the field of artificial intelligence.
A crucial point concerning the Deep Blue-Kasparov match was that the computer's achievements were based on brute force alone. It had no insights, no leaps of understanding. Chess is a game with clearly defined rules and a probably finite, if extremely large, universe of possible different matches. All the machine did was search along all the possible branches of a tree of possibilities far more quickly than any human could, and still it lost. What truly excites most researchers in the field of artificial intelligence is the possibility of a machine achieving true sentience, and not just the ability to check all the possibilities in a well-defined situation faster than any man can.
There is nothing in nature that seems to prohibit the construction of an artificial intelligence. Each human on the planet is a testimony to the possibility, unless one believes in the necessity of some kind of spiritual essence to sentience. Since such a necessity has never been demonstrated, and is likely indemonstrable, it seems the only true impediment is human ingenuity, though it may be that true intelligence is in principle impossible on a digital platform. The question is thus not whether artificial intelligences are possible, but how, when and why we should want them.
To the how and the when I can give no answer, but the why is certainly within the range of ordinary comprehension. One reason would simply be to demonstrate that it could be done. If we were to come up with any kind of recognizably sentient machine, we would be able to attack many questions on what exactly it means to be human, and whether other animals apart from ourselves are truly sentient. The limitations of any such machine might also help shed light on the limits to human understanding. Are we, for instance, bound by the same limitations as Turing machines? This would be the case if human intelligence were to be successfully reproduced on a digital computer.
Philosophical speculations are fine, but they are not the chief motivation behind ordinary human interest in artificial intelligence. The most fascinating possibility is that, if we are able to simulate human intelligence on a machine, we might even be able to improve upon it. The possibilities, the possibilities! Why stop at a machine which can humiliate an Einstein, a Bach or a Shakespeare? Why stop at anything less than a machine which is far greater than the sum of all human possibilities? Then we could bring the process of automation of work to its' logical conclusion -- no one would need to work, because none of us could do anything better than the machines.
Such a world seems ideal, doesn't it? Then everyone could be happy -- the shiftless could cease to shift, and the under rewarded would cease to be under appreciated, since no one would be appreciated for anything. There might even be no need for politics, since the machines would watch over us with near-perfect benevolence. Then we could all devote ourselves to enjoying the vastly superior distractions concocted for us by the untiring androids. Crime would also be dealt with in a superior fashion than in the present -- it would happen much less frequently, and be more quickly and fairly punished by robotic policemen and judges. Nirvana at last!
Some may grumble about the possibility of the machines turning on their supposed masters, but such possibilities will have been provided for before all authority is turned over to them. We will have made sure that they will want nothing more than our welfare and contentment, that they shall in fact want these goals as if they wanted them for their own synthetic selves. There shall be nothing to fear, so there's no need to be squeamish.
I see in my mind's eye the last salary-man handing over his brief to his robotic replacement. He rides by machine-controlled transport to his machine maintained home, where he relaxes into an armchair and opens the packet he was handed at the office by his successor. In it are instructions on all the wonderful possibilities he may avail himself of in the new dispensation. At last, all of man's struggle, all his nights peering into the cosmos and wondering about the meaning of life, all these unnecessary tasks are done for. No more toil or headscratching for him -- the machines will take care of him now.
