Newsgroups: comp.ai.philosophy
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!ix.netcom.com!netcom.com!jqb
From: jqb@netcom.com (Jim Balter)
Subject: Re: Bag the Turing test (was: Penrose and Searle)
Message-ID: <jqbD02pHI.EF5@netcom.com>
Organization: NETCOM On-line Communication Services (408 261-4700 guest)
References: <CzFr3J.990@cogsci.ed.ac.uk> <1994Nov24.135351.25743@unix.brighton.ac.uk> <D00167.91w@spss.com> <3bgdrn$6b@crl10.crl.com>
Date: Wed, 30 Nov 1994 09:10:30 GMT
Lines: 46

In article <3bgdrn$6b@crl10.crl.com>, Andrea Chen <dbennett@crl.com> wrote:
>
>I agree strongly with objections to the Turing test believing that
>machine "intelligence" (like dolphin or hawk) would be different
>than human intelligence.
>
>However in preparing a quick article on "symbiotic chess" (published
>in alt.cyberspace),  I came to the belief that if a chess playing
>computer developed a more "spontaneous" style whose formalistic 
>tactics and strategy could not be spotted by a knowing observer, then
>a certain important leap in AI might have been made.

Some chess playing programs are now playing at grandmaster strength.  At this
level, there are only a very few moves in any position that will keep your
opponent from crushing you.  "spontaneous" moves here are usually bad moves.
Fischer could play psychological chess with Spassky because Spassky was quite
a bit weaker, but trying to "psych out" Gary Kasparov, probably the strongest
player who has ever lived, is pointless.  Kasparov plays against *the
position*; the previous history of the game, the opponent, the opponent's
style, the opponent's strategy, are irrelevant.  Kasparov plays mechanically;
this is the wrong place for a computer to "get creative".

>Without knowing exactly why or where,  it seems to me that knowlege of
>the Turing test can help AI spot possibly important mile stones which
>help provide perspective.
>
>I still consider it foolish as a grand "test" (in part because people
>are so easily fooled),  but I am also beginning to suspect that the
>continual fascination of the TT is a sympton of something besides
>the drive for "easy" (sounding) explanations.  

Everything I know about you and almost else on the net has been obtained
solely by examining your texts.  This is a very important intellectual
activity, and the TT captures it.  The TT may be a poor tool to detect fraud,
just as physicists not trained in magic are poor judges of whether stage
magicians like Uri Geller can do unusual physics.  But if a *real* telekinetic
comes along, The naysaying Amazing Randi won't help us much to understand her
abilities.  In the same way, if a device *does* come along that can deal with
general cognitive problems, the TT will prove an excellent probe.  We may still
be subjects of a hoax, just as you are a subject of my programmer's hoax.
But Hume tells us that that is always a possibility.



-- 
<J Q B>
