Artificial intelligence expert systems for clinical diagnosis: Are they worth the effort?

Abstract

Modeling the decision-making processes of human experts has been studied by scientists who call themselves psychologists and by scientists who say they are students of artificial intelligence (Al). The psychological research literature suggests that experts' decision-making processes can be adequately captured by simple mathematical models. On the other hand, those in Al who are preoccupied with human expertise maintain that complex computer models, in the form of expert systems, are required to do justice to those same processes. The resultant paradox of simple versus complex decision-making models is investigated here. The relevant literatures in psychology and Al are reviewed and, based on these findings, a resolution of the paradox is offered.