On Wed, 08 Sep 2004 02:01:10 EDT, "Byron L. Sonne" said: > I'm just waiting for all the cheesy AI fanboys to start yelling at me > now, but then again, they'd probably be the same kind of clowns that > think passing the Turing Test would mean possessing intelligence(2). > Shit man, there's been mornings where I'd probably fail the Turing Test ;) And that last sentence is *exactly* why Turing phrased the test the way he did. Remember that the *flip* side of "the computer trying to be human" is "the human is trying to prove they're not just another computer". Turing's original paper on the subject (http://www.abelard.org/turpap/turpap.htm) is quite readable, and worth the effort. Consider - we (presumably) consider Stephen Hawking both "sentient" and "intelligent" based on what comes out via the turbocharged TI Speak-and-Spell that's his entire link to the outside world. Now, if the *same* output had come out of *another* TI Speak-and-spell, but was (unknown to you) the output of a very large AI rather than coming from liveware, would you ascribe sentience and/or intelligence to it? How would you distinguish Hawking's output from the AI's? And what are the implications if you can't tell the difference? (Turing's take on it is that it doesn't matter if the output is coming from Hawking, an AI program, or a Chinese Room, or whatever - if it acts intelligent or sentient, it *is* for all practical purposes....) (And for the record, I'm not an AI fanboy, nor do I think that passing the Turing Test proves "intelligence" - merely that if it passes, it's reasonable to proceed *as if* it in fact had intelligence... But then, most people don't realize that the flip side of Descarte's famous "Cogito, ergo sum" is that it's about as far as you can get with pure logic. At that point, you get to either choose solipsism or proceed *as if* what you perceived was at least semi-congruent to some "reality"....)
Attachment:
pgp00016.pgp
Description: PGP signature