HEURISTIC KNOWLEDGE – 21061
Can computers and robots learn from their mistakes, and improve their knowledge by trial and error? Is it possible for a machine, or a network of machines, to evolve on its own? Some artificial intelligence (AI) researchers believe so. The existence of heuristic knowledge, or the ability of a machine to become smarter based on its real-world experience— literally learning from its own mistakes—is a classical characteristic of true AI.
Suppose a powerful computer is developed that can evolve to higher and higher levels of knowledge. Imagine that, one day after the machine has been put into operation, it has intelligence equivalent to that of a 10-year old human; and after two days, it is as smart (in a rudimentary sense) as a 20-year-old. Suppose that after three days, the machine has knowledge equivalent to that of a 30-year-old research engineer. Suppose that more and more memory is added, so that the limit of knowledge is determined only by the speed of the microprocessor. What will such a computer be like after a month? Will it have the knowledge of a 300-year-old person (if people lived that long)? Moreover, does an ever-increasing level of intelligence imply that a machine can also become “wise”?
Machine knowledge becomes far more powerful when computers are given the ability to control mechanical devices, as is the case with autonomous robots. Intelligence and knowledge alone cannot build cars, bridges, aircraft, and rockets.Perhaps dolphins are as smart as people, but these marine mammals lack hands and fingers with which to manipulate things. A computerized robot is to a computer as a human being is to a dolphin.
Can computers ever become smarter than, and perhaps more powerful than, their makers? Some scientists are concerned that AI will be misused, or that it could evolve on its own with unintended, unexpected, and unpleasant results. Other researchers believe that the potential benefits of ever-increasing machine knowledge will always outweigh the potential dangers, and that we can always pull the plug if things get out of control.