Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!elroy.jpl.nasa.gov!lll-winken!decwrl!shelby!rutgers!bellcore!pyuxp!pyuxe!nvuxr!ccw From: ccw@nvuxr.UUCP (christopher wood) Newsgroups: comp.society.futures Subject: Re: prognostications about expert systems (was: Looking Backward) Message-ID: <1581@nvuxr.UUCP> Date: 15 Mar 90 17:43:48 GMT References: <53291@bbn.COM> <1578@nvuxr.UUCP> Reply-To: ccw@nvuxr.UUCP (22456-christopher wood) Organization: Bell Communications Research, NJ Lines: 38 In article tom@amara.uucp (Tom Doehne) writes: > > Chris Wood and JoSH have overlooked a serious risk of using expert >systems to substitute for human judgement: we learn better judgement >by exercising it. I think along the same lines; I just didn't express it very well. There is a difference between using judgement and using rules. Expert Systems use rules. Beurocrats use (or abuse, as the case may be) rules mostly, and very rarely judgement. My first response to JoSH included "sometimes you gotta break the rules", a burger king slogan, but a truism in human judgement. Judgement is not applying rules, it is deciding when the rules available should be applied, when new rules that are beneficial should be created, and so forth. There will always be room for judgement, because others are always reacting to rules to better their positions. (in fact, this is an analysis that lawmakers all to often ignore - but that's another subject). [trimmed; the difference between rules and judgement explains most of Tom's concerns] > My own guess is that expert systems will be useful in domains >which no longer progress rapidly, or that they will come with >substantial teaching and explaining facilities and will be used as >assistants and training tools but not depended upon. >Tom Doehne I'll agree with the former, and say hmmm, maybe to your latter point. -- Chris Wood Bellcore ...!bellcore!nvuxr!ccw or nvuxr!ccw@bellcore.bellcore.com