Author: Ralf Elvsén
Date: 09:46:00 06/29/01
Go up one level in this thread
On June 29, 2001 at 11:21:22, Graham Laight wrote: >On June 28, 2001 at 09:08:48, Ralf Elvsén wrote: > >>My main problem with the Turing test is that a member of, say, a >>technologically advanced species from outer space wouldn't pass it. >> >>Ralf > >Welcome to the world of intelligence testing! > >IQ tests, while they have some predictive power WITHIN cultures, are no good >when used BETWEEN cultures. I get annoyed even with USA IQ tests, because some >of the questions are culture based, and therefore put me at a disadvantage (I'm >English). When I've tried giving English IQ tests to intelligent French people >(who speak good English), their scores were surprisingly low - too low for >credibility. > >They're certainly not supposed to work between species. > >The turing test is designed specifically for computers. If, upon being visited >by aliens, we decide that our top priority is to assess their intelligence level >relative to ours (which would have been a good idea for the occupants of the >host planet in the film "Morons from outer space"), then I suggest we design >appropriate tests for that situation at that time. > >-g It seems I didn't get my point through. If a computer can't be intelligent just because it can't mimic specifically human characteristics, the alien-example shows the weakness in the test, since it would be obviously intelligent. The Turing test should be like this: if you can't tell whether it's a computer or an alien of the above mentioned kind you are communicating with, then the computer is intelligent. This will give the computer the choise of not having to emulate (human) emotions or give opinions on poetry (unless it claims to be a vogon :) . Ralf
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.