Path: utzoo!attcan!uunet!zaphod.mps.ohio-state.edu!mips!dimacs.rutgers.edu!aramis.rutgers.edu!athos.rutgers.edu!christian From: djohnson@ucsd.edu (Darin Johnson) Newsgroups: soc.religion.christian Subject: Re: Who will teach morals to computers? Message-ID: Date: 22 Oct 90 06:10:35 GMT Sender: hedrick@athos.rutgers.edu Lines: 55 Approved: christian@aramis.rutgers.edu In article , mclarke@ac.dal.ca writes: > There are tow evolving technologies which will gradually change the > way society makes day to day choices. They are artificial intelligence > and parallel processing. Essentailly this means that within about 20 years > computer technology should be able to emulate the human reasoning process.. 20 years ago many people predicted this would be the current state of affairs now... Most researchers now would rather pick a name different from 'artificial intelligence', since the term leads many people to assume that the goal of this research is to produce the sort of self- aware computer common in science fiction. Instead the major emphasis is on producing computers with more intelligent behavior than they have now. For example, a program that plays chess would have more intelligent behavior by exploring a few good moves than by exhaustively searching all possible moves. Of course, the word "intelligent" itself has very vague meanings. In the past, being able to rapidly solve some math problems was considered 'intelligent', however after computers were able to solve these problems the definition changed. One common definition of intelligence in the field is "anything a human does better than a computer". > When the technologies mature corporate and government choices can be made by > computers. They will be more efficient, faster, more accurate and much > cheaper. Professional jobs will be done by computer. This includes > medical, legal, etc. positions. > > The question is " WHO is going to teach the computers human values, morals, > etc." Christian bodies had better start investing $$$ in research now to > avoid future crisis. A computer that needs morals is so very distant and improbable that investment now is useless (if it ever happens, who's going to remember who gave who money 200 years ago?) If man-made machines ever get this advanced, they will still be the tools of who ever uses them. Ethical training should start with the users first. (personally, I don't think a machine generally accepted as 'thinking' will ever exist) -- Darin Johnson djohnson@ucsd.edu [Certainly we are some time away from a system that is as autonomous in as large a range of activity as a human being. I refuse to speculate whether such a thing will ever exist. However we already have systems that are goal-directed, and some work is going into a bit more autonomy in choosing goals. We already have to start thinking about ethical implications of how computers are used. (Those interested in such issues should be reading comp.risks.) I think rather than having one day when we suddenly realize that computers are intelligent and we have to start teaching them ethics, the nature of the ethical issues involved with their use will slowly change as they become more autonomous and use higher-level goals. If we're lucky, the changes will happen slowly enough that we don't get taken by surprise. --clh]