Artificial intelligence

(By J. Hilts)

Eventually I got the chance to go to California and to visit the Stanford Artificial Intelligence Laboratory. On a hot June afternoon I met the lab’s director, John McCarthy. I had driven up from the main campus of Stanford University to the outpost in the hills. He was late, so I waited in his office. It was the head of a long snake of a building that sat coiled on the hot Mil-top. Two walls of the office were glass, and through them I could see the Mils outside, which were the color of straw. The short, yellow bristles of grass made the hills look like the scalp of a marine recruit with the wiry dark hair of bushes and trees shaved off, the bumps and scars of contour were visible. The few trees out the window were eucalyptuses, and they looked dusty and dry as fence posts.

John McCarthy’s appearance, when he finally strode into the office, struck me as extraordinary. He is about average height, five feet nine inches. His build is average, with a little age trying to collect itself around his middle. But his hair encircles his head and his face with a great cloud of silver needles. Amid the prickly gray mist his eyes are two dark rocks.

Our first meeting actually consisted of several conversations, between his bouts of work. I remember most clearly one moment, a pause between talks. There is a long wooden table in his office, and I recall the form of Professor McCarthy seated before it. His body was hunched slightly in the shoulders, held motionless, and his eyes were rapt. A small screen and keyboard were in front of him. The machine was in a little clearing amidst a jungle of papers and ragged envelopes. I had come in and sat down, but for a moment my presence was immaterial, a shade at the rim of the consciousness. He continued staring into the screen. I rec­ognized this sort of catatonia. Scientists (as do writers and artists) wander into the paths in the background of their work and cannot find their way back immediately. I didn’t interrupt him.

There are about three million computers in use in the world now. But not millions or thousands, or even hundreds of them are dedicated to the sophisticated work of artificial intelligence. Though there has been much celebration of the coming of the computer revolution, it can hardly be said to apply to our current use of these machines: They do little beyond arithmetic and alphabetical sorting. In practice they are no more than automated filing systems with central controls, and still the chief task they are assigned around the world is to keep track of company payrolls. The promise of computing — "the steam engine applied to the mind" as one professor of computing put it — still remains largely unrealized.

The one tiny academic discipline in which the limits of computers are being tested is the field of artificial or machine intelligence. Of the hundreds of thousands of com­puter programmers in the nation, only a few hundred have devoted themselves to the question of what computers are finally capable of, asking whether the old science fiction saw about brains and computers being equivalent is, in fact, actually true. It has been said for fifty years that computers are "giant brains" and that the human mind is merely a "meat computer." In a slightly dif­ferent form, the same idea has been expressed.

Within this small, exotic field, John McCarthy is one of the three or four people who have contributed most. As I sat waiting for McCarthy to finish, I could see him blink a few times and retrieve his thought from the screen before him. I could see he was beginning to recover. He rubbed his eyes beneath black frame glasses.

He began by saying that in artificial intelligence the object is to find out what intellectual activities computers can be made to carry out. He is rather certain that an intelligence smarter than a human being can be built. From time to time, journalists who discover the existence of his laboratory call up McCarthy to ask him about such robots: "Can they be as smart as people?" McCarthy smiles, "No. That is one of the science fiction fantasies, that robots will be just as smart as humans, but no smarter. Robots will be smarter, because all you have to do is to get the next-generation computer, build it twice as big, run it a hundred times as fast, and then it won’t be just as smart anymore... The field of artificial intelligence is a collective souls people attempt to create such machines. There are now about three hundred souls in the United States and perhaps another two hundred in the rest of the world working to make them."

Дата добавления: 2015-07-16