From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.ecf!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!unix.cis.pitt.edu!kpsst1 Mon Jan  6 10:30:17 EST 1992
Article 2466 of comp.ai.philosophy:
Xref: newshub.ccs.yorku.ca sci.philosophy.tech:1675 sci.logic:711 sci.math:5432 comp.ai.philosophy:2466
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.ecf!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!unix.cis.pitt.edu!kpsst1
>From: kpsst1@unix.cis.pitt.edu (Kent P Stiegler)
Newsgroups: sci.philosophy.tech,sci.logic,sci.math,comp.ai.philosophy
Subject: Re: Penrose on Man vs. Machine
Keywords: analytic arguments, reflection principle, standard model
Message-ID: <200432@unix.cis.pitt.edu>
Date: 1 Jan 92 21:10:31 GMT
References: <1991Dec23.213632.18047@cambridge.oracorp.com> <1991Dec23.190337.6899@husc3.harvard.edu> <1991Dec27.014138.3071@grayhawk.rent.com>
Organization: University of Pittsburgh
Lines: 58

Just a few comments from a humble beginner-philosopher:

In article <1991Dec27.014138.3071@grayhawk.rent.com> siproj@grayhawk.rent.com (D. R. Arthur) writes:
>Since man is inherently creative, and an AI should by definition by designed
                                             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>initially to be more advanced in its thinking process than a human, to have
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>real purpose since real humans are much more easily reproduced and less 
>expensive than an AI would be to begin with.

In all my Introduction To AI courses, the definition never included anything
that would suggest an AI would be more advanced than a human.  I was taught
(Haugeland)  that:
1. Strong thesis: We are computers.  Appropriatly programmed computers
literally have cignitive states... AI *is* genuine intelligence in the
same sense that human inteliigence is genuine intelligence (not a fake
immitation)...
2. Weak theses: anything humans can do, computers can seem/appear to do...
when tested for cognitive abilities, we cannot differentiate between a computer
and a human...

As for the rest of your statement, the cost/easy of production/purpose of
generating AI is not part of the philosophical or logical argument at hand ... 
is it?

>Does anyone have a list of researches that propose AI development that excedes
>the mind model of humans?
>
>Does anyone or no one agree with the above proposition, that an AI needs to
>be more advanced in mind, than humans, to have real purpose?

Depends on what *kind* of advancements you mean.  If you mean that AI
would be an answer to the result of another 400 (pick a number) years of
human evolution, that's a reasonable expectation.  The *goal* however is
to equal us at our present stage... besides, I would presume that the 
basic enigmas and contraditions (machines that think/where is the soul and
mind/etc) would be the same BASIC problems for AI now and in future evolutions.
Isn't the problem hard enough already without wanting AI to be even better 
than us?

A friend suggested to me that since the average person is not a highly 
skilled or brainy individual (rocket scientist, doctor, whatever), 
accomplishing an AI equivalent of a normal person would be rather boring.
I asked her if she would consider it a failure if I created an AI that
tested as severly retarded by human standards.  She said yes.  While it
would be disappointing empirically, it would be a triumph in the 
theoretical sense.  Perhaps even moreso, since our understanding of
what causes someone to be retarded is hardly clear.  Indeed, it seems 
like a likely result for a first-attempt at AI.



////////////////////////////////////////////////////////////////////////////////
/\ Gary S. Wachs (412) 363-4242       \/        University of Pittsburgh.      /
/\ U.S.Mail: 1353 Cordova Road,       \/ "Energy, matter, entropy & antimatter /
/\ Pittsburgh. PA 15206-1430.         \/ sat down to resolve their differences./
/\ Internet: kpsst1@unix.cis.pitt.edu \/  Alas, they agreed to disagree."      /
\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\


