Copyright 1999 by Donald R. Tveter, http://www.dontveter.com, commercial use is prohibited. This material cannot be quoted at length or posted elsewhere on the net or included in CD ROM collections. Short quotations are permitted provided proper attribution is given.

My recommendation: don't worry too much about the definitions at this point in time. First we need to find out EXACTLY what people are doing when they show intelligence, are thinking, are understanding, ... and then the definitions will be obvious. The problem now is that many people have already assumed different theories to be TRUE and all the disputes occur because of which theory they hold to be dear. For instance, some people currently BELIEVE that thinking is a form of digital computing and they produce definitions of these terms based on that BELIEF. Other people BELIEVE that thinking is a form of neural computing that cannot be exactly duplicated by a digital computer, this leads to other definitions for these terms. Other people BELIEVE that thinking is a form of quantum computing and this leads to yet other definitions. Another BELIEF is that there is a human soul and no machine will be able to duplicate what a soul can do thus machines will never be equivalent to people. And so on and so on. This introduces the next question in this document where the issue is to determine when the job is finished, that is when will we know EXACTLY what people are doing so we can compare it with what the machines do?

Whenever I've mentioned that we need to find out EXACTLY what people are doing I get the criticism that that is not necessary because we really want and should be able to build intelligent systems without knowing how people operate. That is, we don't need machines as flawed, emotional and unpredictable as human beings, we want and need better than that. My counterarguments are first that it would help a lot to know how people REALLY DO operate so we can do much the same thing in an artificial system, the artificial system would not have to operate EXACTLY as people do, but the clues you get are really priceless. I suspect too, that if we do manage to create a "better" artificial intelligence, without human emotions and human flaws then its flaw will be that it cannot understand much of the world including the people who created it thus while it would be valuable in some ways it will be a failure in other ways.

Second if we make machines better in some ways than people that will certainly be fine but if we're going to answer the "What is" questions we do need to know EXACTLY what people are doing so we can determine if what the machine is doing is so close that we should apply the words, "understanding", "thinking", "creativity" and so on to what the machine does.