Uncertain Times
Technology shapes the times. Eighty years ago, the world had entered the
atomic age. At first, only the U.S. had
the bomb, but within a decade the “atomic club” included Britain, France, and (crucially)
the Soviet Union. I was born and grew up
under the cloud of nuclear worry: Would the missiles fly? What happens to human life if they do? More decades passed, with more countries
joining the club: China, Israel, North Korea, etc. Only God (and the CIA?) knows which heads of
state have the power to kill millions.
Worries about thermonuclear weapons overshadowed
the world I grew up in. International
leaders had to learn to talk with each other; when they failed to talk (as in
October 1962) ordinary people cowered before the threat of war. Political leaders concerned themselves with many
diverse issues, but no one could escape the question of preventing war. Peace through negotiation, peace through
strength, peace through prosperity, peace through advanced technology (anti-ballistic
missiles), and so on. Only madmen would straightforwardly
advocate war.
Questions of nuclear security
continue. Mr. Trump says Iran must never
have a nuclear weapon. The judgment here
is, I suppose, that no country dominated by certain versions of Islam can be
trusted with weapons of mass destruction.
Maybe that’s true. But it might
be helpful to remember that communist governments have held nukes for decades
without using them. In the 1950s and
1960s, many voices said that godless Marxists could not be trusted with the
bomb. Gradually, we came to rely on
mutually assured destruction. Ideological
opponents could learn to “coexist” (language of the 1960s) if they believed the
other side believed that war would end civilization.
We’ve learned to live with the bomb,
even if we haven’t learned to love it. (Remember
Dr. Strangelove) That means living with anxiety, because some day some
man or woman may press a button.
In this decade, we’ve become aware
of a new technology that may be just as anxiety-producing as H-bombs. Our computers have been growing more powerful
at an exponential rate. According to
Moore’s law, which is really an observation of recent history, not a natural
law like Boyle’s law of gases, says that the number of transistors in an
integrated circuit doubles every two years.
In 1975, Gordon Moore speculated that this rapid increase in calculating
power would continue for another ten years.
It has continued for fifty years, and tech companies have built empires
in the expectation that it will continue.
Much of our post-industrial prosperity has come from the amazing growth
in computing power. Your hand-held phone
contains a camera and a light and connects you to the Internet.
But now we learn about Chat GPT—and dozens
of other programs that can write to us (and for us). Do the computers also think, or do they “just
run programs”? (Remember Short
Circuit) Tech experts and philosophers are working on that question. Interestingly, some of the big tech companies
have started hiring Philosophy PhDs to help them think about their products. We know that AI is coming.
Are intelligent, autonomous
computers a threat? Already we have
self-driving cars (which I predicted in 2015), and we laugh when we read of a group
of Waymo taxis lost in a cul-de-sac. We
assume that all self-piloting vehicles will be programed to protect passengers
and other human beings. (Remember Isaac Asimov’s
first law of robotics.)
We’re pretty confident that medical
diagnostic programs, translation programs, vehicle piloting programs, and many
other complex computer systems will be benign.
We have some worry about energy management. If there is not enough power to go around—in a
winter storm, let’s say, when the grid loses some capacity because of wind, and
it’s really cold—how will a computer prioritize energy distribution?
Of course, we think that’s
solvable. But then we worry about military
uses of AI. The long history of warfare suggests
that speed is often crucial to victory.
Of course, it’s not the only important factor, but smaller armies have
often defeated more powerful but slower foes.
In the 21st century, speed of decision making has become
crucial. Increasingly, we program our
responses into computers, because AI can respond much faster than
soldiers. To win battles, we program
computers to respond automatically.
When does “automatically” become “autonomously”? We worry about a computer that decides not
just when to fire but whether to fire.
In science fiction terror stories,
the computer decides to eliminate us.
This is the computer apocalypse, which has already achieved “legend”
status. I’m confident I could find
several such stories. (How? By a computer-aided Internet search, naturally.)
But our worries go further than
that. The autonomous program may decide
that a certain weakness in the enemy’s defenses should be exploited—and the
missiles fly. How do we use computers to
fight without letting them decide whether to fight?
We have other worries about AI,
unrelated to wars or bombs. Some people
have used AI technology to produce “revenge porn.” I fully expect to see political ads containing
falsified images of politicians in coming years. Our social and political discourse has already
been harmed by the way we retreat into Internet ghettos; AI will make it easier
and more emotionally satisfying to ignore the humanity of the other side.
Apparently, there is profit in programming computers to call phone
numbers (selected randomly or by algorithm) to sell insurance or home repairs
or many other things. I have made it my policy
to never buy a product advertised this way; I insist on talking to a real
person.
We live in uncertain times. To
some degree that has always been true.
But my lifetime, spanning the atomic age and the digital age, seems to
be especially uncertain. The fact that
we have survived eighty years of the atomic age may give hope that we can meet
the challenges of the digital age.