I am the subject of the article "Artificial Stupidity." Generally I was quite pleased with the article.
However there are a few comments regarding the article and reader's responses that I would like to make.
1. I am not, and was not, a "control freak" micromanaging Loebner Prize contests.
True, I was and am a "control freak" on the topic that the contest must be held annually. (Actually, I would say the term "control freak" is much too mild -- perhaps "rabid" is more appropriate.) However, I offered the prize on the condition that a contest would be held annually. The Loebner Prize Committee, of which Dennett eventually became chairman, unanimously accepted the requirement that the contest be held annually. I would say that requiring the contest to be held annually is not "micromanaging." It seems to me to be the most macro decision one could make.
Given that the contest was to be held in any year, I sometimes provided suggestions (only suggestions) to that year's host. These they promptly ignored and I did not protest or argue. You might want to check with Dave Powers of Flinders U, James Moor of Dartmouth, Sabiha Foster of the London Science Museum, and even Daniel Dennett at Tufts, each of whom directed one or more contests. Each of them was free to run the competition as he or she saw fit, with no directions, instructions or supervision from me. Had I "micromanaged" the contests, the 2002 contest in Atlanta would not have been so disorganized.
2. The Loebner Prize Contest was held twice in Australia under the auspices of Flinders U, and Dr. Powers offered to host the contest a third time. Not everyone said "never again."
3. Sundman thinks me a fool for stating that I could hold the contest in my apartment. I have personally directed three contests and observed nine others; Mr. Sundman has directed none. I know from personal experience whereof I speak. The Turing Test may be a deep topic but the test itself is not particularly difficult to implement physically. I could (and actually might) hold a contest in my apartment with no great difficulty.
To those readers who claim that the Turing Test is not a good measure of intelligence I would reply that it is one measure of intelligence. I have no objections, philosophical, moral, or personal to other measures or tests. It is just that I have an interest in the Turing Test, not the other tests. If the Turing Test shows an entity to be "intelligent" but the others don't, or vice versa, we may have a problem. On the other hand, I think it likely that any computer entity that passes a Turing Test will likely do quite well on other measures of intelligence also. It will be interesting to see if this is so.
To those readers who claim that the Turing Test is a valid measure of intelligence, but that the Loebner Prize contest is not a good Turing Test, I ask "why not?" In what manner is the Loebner Prize contest, as currently held, different from the exercise that Turing described? Note that the horrible rule that conversations have to be restricted has been removed.
-- Hugh Loebner
Shares