Issue No. 04 - August (1995 vol. 10)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/64.403961
<p>The joint cognitive system approach to problem solving works especially well in creative situations, such as design, that have no single "correct" solution. Our experience in providing expert systems support to a biotechnical equipment manufacturer argues for a shift from classical, problem-solving expert systems to more cooperative, advice-giving systems.</p> <p>Expert systems represent applied artificial intelligence in its most successful form. Yet while such systems have solved many difficult problems in a wide variety of settings, talk persists about their failure. The major problem, it would appear, is poor user acceptance. Too often, users complain that such systems solve problems that don't need solving, or don't match the way people actually perform their work. Consequently, upon delivery some expert systems simply find no use among their presumed users. Our experience shows that closer attention to the actual difficulties presented by the tasks users want supported offers one remedy.</p>
G. Forslund, "Toward Cooperative Advice-Giving Systems," in IEEE Intelligent Systems, vol. 10, no. , pp. 56-62, 1995.