Path: utzoo!attcan!uunet!bu.edu!dartvax!eleazar.dartmouth.edu!phil
From: phil@eleazar.dartmouth.edu (Phil Bogle)
Newsgroups: comp.ai.philosophy
Subject: Re: Testing for machine consciousness
Summary: Does consciousness even matter?
Keywords: consciousness
Message-ID: <25036@dartvax.Dartmouth.EDU>
Date: 10 Oct 90 16:43:00 GMT
References: <7@tdatirv.UUCP> <1990Oct8.120927.8648@canon.co.uk> <21@tdatirv.UUCP>
Sender: news@dartvax.Dartmouth.EDU
Organization: Dartmouth College, Hanover, NH
Lines: 25

There has been a great deal of discussion of consciousness as an emergent
phenomenon and the potential of machines to attain it.  I'd like to
backtrack to ask a fundamental question: why should we care whether a
machine is actually conscious, sentient, "aware", and so forth?   With regards
to intelligence, consciousness as opposed to behavior which merely
_seems_ conscious makes no difference whatsoever.  Would AI research be
considered a failure if actually did manage to construct Searle's non-conscious
Chinese Room?

  Even from the standpoint of ethics, would you feel no shame
destroying an intelligence capable of creating elegant and beautiful
ideas, merely because it wasn't "sentient" (especially if it gave every
appearance of being so.)

   I'm not saying I would freely abandon my self-awareness.  I am
suggesting that Searle and others have their priorities messed up.  The real
question should not be "Can we create intelligence without consciousness?", but
"Could we create intelligence without having some kind of emergent, higher level
structure?" It would be very distressing if a machine like Searle's CR
managed to fake it's way through the Turing test using only a set of
disconnected, low-level rules-- not because the machine isn't conscious,
but because it has none of the complex, emergent structure we associate
with intelligence.  Other than that point, however, I can't see how Searle's
argument should influence AI researchers in any way.  The field, after all, is
AI, not AC.