I find the strong AI hypothesis interesting - ie the question of
whether it is possible in principle for a computer to be conscious or
self aware.
I agree that some of the arguments against strong AI are very lame.
For example Searle says a simulation of mind is not the same thing as
mind in the same way that a simulation of a hurricane is not the same
thing as a hurricane. But if mind is computation, isn't a simulation
of a computation just an isomorphic computation?
My guess is that the strong AI hypothesis is true but the
repercussions are quite astounding. If mind is computable then it
follows that there exists Turing Machines representing worlds with
self aware substructures (SASs) as complex as ourselves. Presumably
many SASs would be tricked into believing in a distinction between
platonic existence and their apparent physical existence. By
application of Occam's Razor I would argue that the (mystical)
postulate of our own physical existence should be dropped in favour of
a less presumptuous platonic existence. That's ironic because most
people would say belief in physical existence is less mystical than
belief in platonic existence - but each to their own!
The best argument against strong AI that I've seen is the White Rabbit
Paradox, which basically says that if we are merely computation, in
the multiverse of all possible mathematical realities or computations
why don't we see bizarre things happen like white rabbits appear out
of nowhere. However there is a possible resolution based on
"probability" (or measure) theories on infinite ensembles. However
to really be taken seriously the researchers in this area need to
explain our laws of physics. I do find their simplified explanations
of some of the bizarre features of Quantum Mechanics interesting - the
Many Worlds Interpretation of QM seems a natural fit with the ultimate
ensemble theory. However I'm not sure whether they even have an
explanation of why Hilbert spaces over complex numbers are so
important.
Anyway I don't think there has been any compelling argument either for
or against strong AI.
Received on Sat Jul 12 2008 - 06:10:35 CEST
