From Stan Hayward
London
It was interesting to see the comments on Deep Blue 2’s use of artificial
intelligence to defeat Garry Kasparov at chess
(“Slaughter on Seventh Avenue”, 7 June, p 26).
Of course, unthinking intelligence is not the prerogative of supercomputers.
It is also common to many very intelligent people who make blunders that are
obvious to less intelligent but more experienced people. Government departments
are a rich source of unthinking intelligence.
The question is: how does one use artificial intelligence to a practical end?
I would like to suggest a new approach to the subject and propose the
development of “artificial stupidity”. The aim of such a machine is not to
suggest the best solution to a problem but the solution that people dealing with
the problem are most likely to come up with.
Advertisement
For example, the government currently pays around £400 a week to
incarcerate absent fathers who cannot afford to pay £40 a week
maintenance. An artificial stupidity machine would calculate that, as single
mothers are increasing at a rate of around 40 000 a year, then the solution is
to build more prisons. This is, of course, one of the government’s
proposals.
Such a machine would arrive at this solution much faster and more cheaply
than a committee. An artificial intelligence machine might suggest putting money
into studying why the divorce rate is so high, and suggesting means of
preventing it. But as the article shows, artificial intelligence works
differently from humans.
