Submitted by: Sarang Gupta (sarang@sarangworld.com)

Blondes and Boors: How Artificial Dumbness Will Pass the Turing Test

English mathematician Alan Turing proposed than a computer could only
be considered truly intelligent if it could converse with us at the
level of a human being. His challenge, the "Turing Test", has sparked
computer scientists to design and implement complex artificial
intelligence programs, some using gigabytes of memory and performing
billions of computations per second.

However, the bulk of humanity is not quite so sophisticated. We are all
familiar with the jocular tale of the student who graduated with highest
honors, even though his vocabulary was limited to the five words, "Yes,
Sir.  You're absolutely correct.". And, if one were to ask the average man
what type of woman he would prefer for a dalliance, the answer could also
doubtless be coded in fewer than a hundred words: the phrases, "right
away, honey", "you're so smart", and "oh, I must be wrong again" spring to
mind almost immediately. These repetitive, limited-vocabulary humans are
trivial to simulate.

And what of those people we meet frequently at outings? Perhaps they are
more verbose, but their single-mindedness limits their subject matter. We
are all familiar with the chatterbox who will spout baseball statistics at
the drop of a hat, and speaks of nothing else -- or the woman who brags
incessantly of her children, and hears nothing else. These people, too,
are not difficult to simulate: merely create a program with vast knowledge
on one subject-- and whenever the topic strays from this one subject, have
the program single-mindedly ignore all distractions and continue. 

Simplistic? Yes. Effective? Undoubtedly. One might even argue that a
sophisticated computer program that is able to speak intelligentally on a
variety of subjects would fail the Turing Test-- after all, Reinassance
men are rare in this day and age, and a computer that could simulate one
accurately would draw suspicion. 

And what of practical uses? Proponents of virtual reality decry that
computers are not yet sophisticated enough to accurately model human
behavior. To which I respond, "If one wishes to experience accurate human
behavior, one need only stand on any street corner for 5 minutes. I, for
one, would much rather experience virtual UNreality".

The advent of artificial dumbness will make such products available: The
Virtual Blonde Girlfriend, whose limited vocabulary will have fewer than
100 words, will in no way be any less popular than her human counterpart
(with perhaps an optional zero-word model for those who feel a blonde's
silence should be as golden as her hair). The lack of sophisticated
conversation could easily be advertised as a feature, though I suspect
most of the mass market for this product is unlikely to even notice this
lack. 

And for those who wish to get away from even a virtual girlfriend, the
Virtual Drinking Buddy will provide an endless ream of useless trivia,
while remaining oblivious to the outside world.

But what of those who do seek intelligent conversation, those who want
more in a relationship than a mere trist with an airheaded automaton--
those who value friendship as a sharing, bonding experience, not as a
blathering respite from thought? What will they do? When will
computer-generate reality reach their level of sophistication? When will
artificial intelligence become a reality? Ask yourself these questions
now, because, in 20 years, when 99% of the population is living in virtual
unreality... no one's going to care about the answers.