The
Quest for Intelligence
by Richard Jones
(Up to General
AI)
I have had a keen interest in artificial intelligence which
has been complimented by many modules I have undertaken on my university
course. The following is an article I have written expressing my
views on the field of artificial intelligence. Any views on this
article and comment would be greatly appreciated. Please email
me with them.
Intelligence is a subject that has been debated for thousands of
years and the debate has no signs of finishing. A dictionary definition
of intelligence is "quickness of understanding." Even though this
is a particularly apt definition such a simple clear statement has
fatal flaws within. There are many definitions of intelligence each
with these flaws. These endless flawed definitions prove troublesome
if trying to recreate intelligence.
Attempts to recreate intelligence have been the life's work of
many scientists. Intelligence can be demonstrated in a software
program. However it can have more impact if software is used in
conjunction with robotics. A software program that can add up 20
numbers and display a result very quickly can be seen as exhibiting
some form of intelligence. A robot with an "intelligent" program
that can navigate across a room would be seen as more "intelligent"
by the majority. Hence there has been a significant amount of work
is developing and using "intelligent" robots.
The main problem with developing an intelligent robot is the many
definitions of intelligence. As stated previously a computer program
that can add up twenty numbers may be considered intelligent by
some but not by others. This results in much debate over whether
an intelligent program has been written already.
In society today many people base intelligence on academic achievement.
This is highly unfair. People develop at different speeds and thrive
under different circumstances. One person may peak academically
at fifteen years old, another at twenty-one. Hence their achievements
will be different. Everybody has different skills and trying to
base a person's ability by global standards in unreasonable. The
world today is obsessed with qualifications and obtaining pieces
of paper demonstrating that you can use, for example, MS Windows.
If someone is an Oxford University graduate people automatically
assume they are intelligent. This however does not take into consideration
their social ability and also whether they are just intelligent
in one particular niche. For example someone could graduate with
a First class degree from Oxford University in Computational Mathematics
but not be able to spell very well. It could be questioned whether
they more or less intelligent that someone who obtained a third
class degree in the same subject from a less pretigiuous university
but who can spell much better?
This indicates that we only look at someone's ability in a one
dimensional sense. For example if the Oxford graduate's intelligence
were considered only against a computation mathematics scale then
he/she would be considered very intelligent. If it were considered
against a spelling scale then he/she would be considered less intelligent.
However society does not take into account this one dimensionality
of intelligence.
The proof of this one-dimensional view is how we measure a person's
intelligence. Many believe in the results of the Intelligent Quotient
(IQ). It can be argued there are many types of problem solving questions
but again this is only a niche of what can be considered to be judged
for intelligence. Is musical ability tested by IQ? Mozart is considered
a genius universally but one wonders how would he fare in an IQ
test. However it can be argued is a genius intelligent? They can
be because they can achieve what many others cannot.
So intelligence should be considered multi-dimensional. What dimensions
should be included in the test and how can a comparison be made?
This is a very difficult question. For example it could be argued
that an intelligence test should test knowledge about a person.
Then it could equally be argued that he/she is the most intelligent
person in the world. No one knows more about himself or herself
than the person in question. Many people would argue that knowledge
on a particular person, as one of the dimensions is unfair. However
all subjects for the different dimensions would never be agreed
upon. They could include musical ability, mathematical ability,
social skills, etc. Indeed the list could be endless and
the inclusion and exclusion of subjects could be convincingly argued
by many.
A classic example of how intelligence differs can be seen by considering
three people who in their own right have been very successful. Albert
Einstein was one of the greatest theoretical physicists ever to
live. His theories are considered very important in modern theoretical
physics. However he was dyslexic. Paul Gascoigne used to play for
the England football team. His ability with a football was enormous.
Obviously a lot of practice was needed but he is still considered
to have one of the best football brains of his generation. He could
do things with a football that many could not and has often been
labelled a genius by both the media and the public. Richard Branson
is one of the most successful entrepreneurs of his generation. He
now owns multi-million pound companies which he has built up from
nothing. However academically he was less successful which is shown
by the fact that he did not pass any O� Levels. Many would believe
that not passing any of these qualifications would limit the success
of a person. In this case it has not. The question must be posed
who is more intelligent; Albert Einstein, Paul Gascoigne or Richard
Branson? Before reading this essay most people would probably have
said it was Albert Einstein. This essay has surely clouded the issue
and each of them can be argued as the most intelligent.
Another aspect of intelligence is that of animal intelligence.
Humans consider dolphins to be intelligent. Is it because they are
just friendly? There are stories of dolphins helping boats through
hazardous waters. Could other animals do this if they were as friendly?
Parrots are another animal that we consider intelligent. They can
mimic our talking and understand what we say. In fact many animals
can. For example when you say "sit" to a dog it can (depending on
training) sit. It can link the noise we make with the action of
sitting. Can humans do this though with animals? Can we understand
the noises that animals make? We can often tell if an animal is
in pain but can we tell if an animal is hungry just by its noises?
Does this make dogs, parrots and dolphins more intelligent than
humans then? Again that is debatable.
In conclusion this essay has shown that the quest for intelligence
is by no means trivial. Or is it? Is it trivial to look for something
that there is no answer to? The author's main research area deals
with artificial intelligence and the quest for the intelligent machine.
However it is believed that the goal of an intelligent machine is
a long way off. We cannot even attribute intelligence properly to
each other. The continuing arrogance of humans wanting to always
be superior would limit us from acknowledging a computer's ability.
We pride ourselves in greatness and ability to be rule over our
world. Therefore if a computer achieves a task better than a human
can then this is not seen as intelligent. The classic examples are
the chess-playing computers. Chess is considered a complicated game
and many of the grand masters are considered very intelligent humans
due to its complexity. However when Deep Blue built by IBM beat
Gary Kasparov, instead of hailing the world's first truly intelligent
computer, it was ridiculed and all its faults pointed out. Man's
arrogance does have its advantages. Our power crazy psyche has the
flaw in that it always wants to be in control. It is for this reason
why truly intelligent machines which could "take over the world"
will never happen. Human beings� own flaw could be our saviour.
Please note that the views expressed in this essay
does not necessarily reflect the views of AI Horizon, but only that
of the author cited.