(These are excerpts from my book "Intelligence is not Artificial")
Footnote: Between Fatima and AlphaGo
Paul Nunez in "Brain, Mind, and the Structure of Reality" (2010) distinguishes between Type 1 scientific experiments and Type 2 experiments. Type 1 is an experiment that has been repeated at different locations by different teams and still holds. Type 2 is an experiment that has yielded conflicting results at different laboratories. UFO sightings, levitation tales and exorcisms are not scientific, but many people believe in their claims, and i will call them "Type 3" experiments, experiments that cannot be repeated by other scientists. Too much of Artificial Intelligence falls between Type 2 and Type 3.
News of the feats achieved by machines rapidly propagate worldwide thanks to enthusiastic bloggers and tweeters the same way that news about telepathy and levitation used to spread rapidly worldwide thanks to word of mouth without the slightest requirement of proof. (There are still millions of people who believe that cases of levitation have been documented even though there is no footage and no witness to be found anywhere). The belief in miracles worked the same way: people wanted to believe that a saint had performed a miracle and they transmitted the news to all their
acquaintances in a state of delirious fervor without bothering to doublecheck
the facts and without providing any means to doublecheck the facts (address?
date? who was there? what exactly happened?). The Internet is a much more
powerful tool than the old "word of mouth" system. In fact, i believe
that part of this discussion about machine intelligence is a discussion not
about technology but about the World-wide Web as the most powerful tool ever
invented to spread myths. And part of this discussion about machine
intelligence is a discussion about the fact that 21st century humans want to
believe that super-intelligent machines are coming the same way that people of
previous centuries wanted to believe that magicians existed. The number of
people whose infirmity has been healed after a visit to the sanctuary of
Lourdes is very small (and in all cases one can find a simple medical
explanation) but thousands of highly educated people still visit it when they
get sick, poor or depressed. On 13 October 1917 in Fatima (Portugal) tens of
thousands of people assembled because three shepherd children had been told by
the Virgin Mary (the mother of Jesus) that she would appear at high noon.
Nobody saw anything special (other than the Sun coming out after a rain) but
the word that a miracle had taken place in Fatima spread worldwide. Believe it
or not, that is pretty much what happens today when a fanatical blogger
reports a feat performed by an A.I. software or a robot as a new step towards
the Singularity. People like me who remain skeptical of the news are frowned
upon in the same way that skeptics were looked down upon after Fatima:
"What? You still don't believe that the Virgin Mary appeared to those
children? What is wrong with you?". Which, of course, shifts the burden of
proof on the skeptic who is asked to explain why one would NOT believe in the
miracle (sorry, i meant "in the machine's intelligence") instead of
pressing the inventor/scientist/lab/firm into proving that the miracle/feat has
truly been accomplished and can be repeated at will and that it really did what
bloggers said it did.
"Whenever a new science achieves its first big successes, its enthusiastic acolytes always fancy that all questions are now soluble" (Gilbert Ryle, "The Concept of Mind", 1949, six years before Artificial Intelligence was born).
Back to the Table of Contents
Purchase "Intelligence is not Artificial")