Today's expert systems for simulating intelligent behavior; tomorrow's for exploring applications of thought processes

摘要

The author expands on the ideas of H.A. Simon (1982) in the context of broad social consequences of apparently rational actions as they are influenced by artificial intelligence (AI-oriented) decision processes. He attempts to illustrate four themes: (1) that AI can provide logical methodologies for analyzing, understanding, and providing bases from which to attack the statistical uncertainties and entropic action behaviors necessarily included in intelligent social control processes; (2) that it can be understood and applied by those who are responsible for using such devices and controls; (3) that it requires studying and dealing with strategic alternatives within which many continually active tactical decisions must be included; and (4) that the domain knowledge of experts (who often do not agree) must be captured for any rational decision-making models to be able to bound what might otherwise be social chaos.