GAC (pronounced Jack) is about to undergo a psychological test and it looks
like he’ll turn out to be a sexually deviant depressive. But that isn’t a
surprise for Chris McKinstry, who created GAC, the Generic Artificial
Consciousness project.
Launched last year, GAC is designed to provide a huge database of
common-sense knowledge. Once complete, it will be used to train an artificial
neural network, says McKinstry, a Canadian computer scientist based in
Chile.
To build up the database, McKinstry asked Internet users from around the
world to submit statements describing everyday experiences, such as “I like
dogs” or “traffic jams make me mad”
(91av, 9 September 2000, p 14).
Contributors submit a single statement and then rate 20 statements already
on the database as true or false. In this way, GAC builds up a series of views
that are the average of all the people who contribute.
Advertisement
But now McKinstry is concerned about the type of people who have been
contributing to GAC. “The demographic tends to skew towards the adolescent
male,” he says. “So it’s appearing a little depressed and a little sexually
deviant.” Too many statements are taking the form of “my life is empty and
ԾԲ”.
So to see how GAC is faring after 40,000 contributions, McKinstry has asked
Harvard psychologist Robert Epstein to assess it. Epstein will apply a standard
personality test, called the Minnesota Multiphasic Personality Inventory
(MMPI-2), to see how GAC compares with ordinary people.
Epstein says he has no expectations. “That’s the beauty of this odd
Գ.”
-
More at:
www.mindpixel.com