|The question Can
a machine think? was first seriously asked, and answered,
in the 1950s by the
tragically short-lived mathematical genius
Alan Turing. You will find his famous paper, with many
other goodies, in James Newman's The
World of Mathematics. But
Science Fiction writers have always known the answer,
and the intelligent machine
is one of the most common characters in the
genre. Ambrose Bierce opened his classic short story,
written in 1909, entitled
'Moxon's Master' with the
words: "Are you serious? Do you
really believe that a machine thinks?" And he neatly
evaded the issue with another question:
"Is not a man a machine? You will admit that he thinks or
thinks he thinks."
Now that great universities such
as the Massachusetts Institute of
Technology sponsor Departments
of Artificial Intelligence this old argument is essentially
over; nevertheless, even the most enthusiastic proponent of
AI would hardly claim that any of today's machines show
more than the most rudimentary intelligence.
What is claimed is that if any intellectual activity
can be precisely described, then a machine can, in principle,
be designed to carry it out. There are a lot of hidden ifs
in both parts of that sentence; but many scientists have now
stated flatly that machines more intelligent than men will
exist in the near future. Some have told me
before 2001, but I do not believe them.
And incidentally, those people
who argue that it is obviously impossible for human beings
to create entities more intelligent
than themselves merely demonstrate that
non-thinking is not confined to machines. Much the same
logic would prove that we could never progress from
Stone Age flint to modern
precision tools, for how can any tool make something
better than itself?
It is probably no coincidence
that when he was at MIT, John W. Campbell
got to know Norbert Wiener,
founder of the modern AI theme,
with such stories as 'When
the Atoms Failed' written in
1930 and 'The Last Evolution'
written in 1932. The
title of the second story
sums up a whole category of later tales, perhaps culminating
in Fredric Brown's famous
'Answer' written in
1954. "Is there a God? There is now."
C. Clarke used a Kaypro II
to write his 1982
novel 2010: Odyssey Two.
The novel was submitted for publication via modem from
thinking machines were almost invariably
Frankenstein monsters, out to destroy their creators,
and usually succeeding; Campbell
was one of the first writers to make them not only benevolent
but even noble. Isaac Asimov
took matters further with his Three
Laws of Robotics, an ingenious attempt to lay down
a protocol for man-machine
relationships, and a generator of endless plot ideas. Some
simple-minded readers have assumed that these laws are
indeed Laws, like those of nature, and not merely rules,
akin to Please drive on the left or right. I have had
the First Law thrown in my teeth
as a result of HAL 9000's
mutiny; to which I have replied that, so far, alas, the world's
most sophisticated robots have been designed for the express
purpose of killing people ...
There is, of course, no fundamental
reason why robots should not be programmed to hate or to love.
See Lester del Rey's 'Helen of Loy',
1938. The first generation of thinking machines would
indeed mirror the emotions of their builders. But later generations
would go on to develop emotions perhaps beyond human understanding.
Men always fear what they cannot
understand, often with good reason. The mathematician
Dr. I. J. Good, who has made a special study of the subject,
has remarked: "If we build an
ultra-intelligent machine, we will be
playing with fire. We have
played with fire before, and it has helped to keep the
other animals at bay."
AI does arrive, we shall be the other animals;
and look at what has happened to them.
It will be
Arthur C. Clarke
Ash, Brian, ed. The Visual Encyclopedia of Science Fiction.
New York: Harmony, 1977, p. 181.
by Mark A Carter: Despite Arthur C. Clarke's
vivid imagination, and that of
Stanley Kubrick, when they worked on the
1968 film 2001: A Space
Odyssey, neither foresaw the remarkable size reduction
in computers that would occur between the late
nineteen-sixties and now. Also, they did not foresee
the increase in computational power and speed, as per
Moore's Law, that would occur every
two years. And they could not imagine anything in reality
resembling the positronic brain
envisioned by Isaac Asimov
in the nineteen fifies. But
Science Fiction is predictive.
How many times do I have to say it? And the latest news,
just off the grapevine, is that
IBM has indeed created the first
brain-like cluster of CPUs.
Google, after all,
will need a sophisticated electronic brain in their
autonomous military robots. It should
come as no surprise that much of the funding came from
DARPA. And after
ten years of development, it is finally a reality. Sorry
Arthur, IBM has developed
what you could not imagine in your
wildest dreams. I guess Asimov
was right after all. And so, it begins.