Persuasiveness of expert systems

Abstract. Expert system advice is not always evaluated by examining its contents. Users can be persuaded by expert system advice because they have certain beliefs about advice given by a computer. The experiment in this paper shows that subjects ( n = 84) thought that, given the same argumentation, expert systems are more objective and rational than human advisers. Furthermore, subjects thought a problem was easier when advice on it was said to be given by an expert system while the advice was shown in production rule style. Such beliefs can influence expert system use.