I was listening to a conference presentation in Bangkok the
other day - he said, shamelessly dropping place names. The subject of the
presentation was the use of artificial intelligence (AI) in managing airline
pricing decisions. It’s a fairly specialised topic and you will be glad to hear
that I don’t intend to say any more about it than that. I was struck however by
an illustration that the speaker used to add colour to his remarks. He was
talking about how the AI makes decisions that the human user of the system
cannot understand. Not only that, but the designers of the system cannot
explain how the machine made that particular decision. Just like in Angry
Birds.
This game was a brief sensation a few years ago. If you are the
one person who has never seen it, it is a smart phone app in which the player
uses a virtual catapult to launch improbably-coloured cartoon birds at a
variety of targets. In the game the launching is done by dragging a finger
across the screen and then letting go. The cartoon bird then “flies” in a
parabolic arc just as if it were a real physical object propelled by the
release of elastic tension and moving under the influence of gravity. The
physics of the situation are modelled accurately and somewhere in the
calculations reference must be made to Hooke’s Law of elasticity and Newton’s
Laws of motion and gravitation. And the player of the game knows nothing of
this. She just knows that if she pulls back the spring the bird will fly in a
certain way. If it misses the target she needs to adjust the point of release.
It is an example of a phenomenon of our modern times, a
technology that is used by millions of people without any of them understanding
how it works. This situation has arisen largely over the last twenty-five years
or so and I believe it is having a profound effect on the way society functions.
Probably not in a good way.
Until some time around the 1970s there was no technology in
widespread use that could not be comprehended by a single individual. When I
studied computers in college we started from the electronics – transistors,
diodes, resistors and the like. We were taught how these things functioned at
the physical level and then how they could be put together to make logic gates
and then how these logic gates could be assembled to make computer processors.
Then we were taught how these processors could be controlled using binary codes
input as varying voltages and how these codes could be built up into computer
languages with which we could do useful work like calculating linear
regressions. We were undergraduates and it was reasonable for us to grapple
with computers from the flow of electrons in semiconductors all the way up to
high-level languages making sophisticated calculations.
The same considerations applied to other technologies like radio
and TV sets, motor cars and aircraft. It wasn’t that these were products of
lone artisans working from their cottages, but somewhere along the line there
was a group of people who understood everything about how they worked.
Not any more.
Modern technology is built in layers. In any kind of intelligent
device the base layer is silicon that has been etched at nanometre scales to
create billions of transistors on a scrap of material the size of a fingernail.
Chip fabrication is an exquisitely precise function that requires factories
costing billion of dollars to set up and run. The engineering of those chips is
a highly specialised skill possessed by a small number of people world wide.
The next layer up is the code that allows those chips to do work, responding to
physical inputs and manipulating voltages that eventually translate to useful
outputs. On top of that are the higher-level languages, the networks, the
aggregation and packaging of devices, the operating systems and finally the
development environments in which programmers can create apps like Angry Birds.
Each layer in the process has its own specialists with their own esoteric
skills. At each layer the engineers and programmers typically know nothing
about the internal workings of the layer above or below. At best they know
things along the lines of “If I do this then the machine will respond by doing
that”.
And the end-user? She only knows that if she slides her finger
just so, the cartoon bird will fly across the screen and break the glass in the
greenhouse. The processes by which that happens are completely unknown and if
anyone tried to explain them the response would probably be glassy-eyed
incomprehension.
Angry Birds is just one trivial example of the way that
technology has advanced to a point where it might just as well be magic. This
was expressed by science fiction writer Arthur C Clarke in his third law which
states[1];
“Any sufficiently advanced technology
is indistinguishable from magic.”
I think that we have reached a point in our development at which
Clarke’s third law has real practical consequences.
We now have a generation of people who have grown up with
technologies that do amazing things and literally nobody knows how they do
them. Some people know parts of the story and they also know that colleagues
and friends know other parts. Rational human beings accept the expertise of
their peers and take on trust the pieces for which they do not have direct
understanding. The problem comes with the irrational ones.
If nobody can understand the full functioning of a device then
it becomes possible to believe more or less anything about it. Maybe there are
tiny little birds hiding inside your iPhone for your entertainment. Ludicrous
idea? Of course it is but when the real explanation is too complicated to
comprehend why not believe something simpler? If it were just the workings of
video games then maybe this wouldn’t matter, but it isn’t.
Global climate change is primarily caused by the emission of
certain gasses into the atmosphere. The mechanism is actually quite simple,
provided you have a basic understanding of atomic physics and electromagnetic
radiation. But for the many millions of people without that knowledge it’s a phenomenon
on a par with shooting birds across a screen. And so it is characterised as
“unknowable”, which allows cretins like Nigel Lawson to go on radio and TV and
claim either that it doesn’t exist or that it is entirely an act of God and
beyond human intervention.
During the Brexit debate Michael Gove infamously said that the
British people have had enough of experts and he may well have been right.
People are so accustomed to technology that does things for them as if by magic
that on an unconscious level they believe that everything in their lives is
amenable to magical intervention. Of course it must be possible to walk away
from the world’s biggest and most prosperous trading bloc with nothing to put
in its place and still have £350 million more per week to spend on health care.
The reasons why this is untrue are complex and depend on an understanding of
underlying factors, most of which are opaque to the mythical man in the street.
So-called “experts” are really spoilsports who deny the existence of magic
despite all the evidence of Netflix, Twitter and a doorbell that shows you on
your phone screen who is trying to get in.
At some point however the magic has to stop. As that great
philosopher Montgomery Scott so often remarked
“Ye cannae change the laws of physics Jim”.
Magic isn’t real. Physics is. Climate change is with us already
as the experts have always known it would be. The “laws” of economics are less
well defined than those of physics but on the whole people who spend a lifetime
observing and studying economic phenomena are more likely to be correct than
those who just pull ideas out of their hats following a booze and fags session
down at the Dog and Duck.
The greatest British novelist of the late 20th
century was Iain (M) Banks[2].
In his science fiction writings he created a universe, The Culture, in which
humans and artificial intelligence live side by side and on the whole
rationality prevails. It’s a captivating utopian vision. Some humans display
irrationality – sometimes on a monumental scale, but these are clearly the
aberrations. They are the exceptions that prove the rule and they usually (but
not always) get their comeuppance. I’d like to think that Banks’s vision offers
a distant goal. A way of thinking about how societies might evolve as technology
advances at a rate beyond human ability to keep up. I’d like to think it but
I’m afraid that the descent into magical thinking that we have seen over the
last few years makes me fear that we might not make it.
