ck the brain of a human expert to find out what sorts of things he or she needs to know to make a decision. The designers then work through the "if-then" rules used by the expert to reach a decision based on the information at hand. These rules are then fed into the computer, and the resulting system can duplicate the decision-making processes of the expert in 80 to 90 percent of cases. In the remaining cases, the computer will recognize that it does not have access to the necessary information or expertise. The computer will in fact state that it lacks such needed information and will then call in a human expert to solve the problem ("Smart advice from dumb machines," p. 61).
Patent (1986) points out that most expert systems are considered only an offshoot of AI because their "intelligence" is arguable and because they are limited in scope, which is also why they are successful. Each system deals with only a narrow area of human knowledge and knows nothing outside that narrow range.
...