Martin Ford -- Rise of the robots
===================================

So, which jobs and tasks can be performed by a computer or robot?
Which jobs and occupations are in danger of being replaced by
automation and robots?  Ford spends a good deal of time trying to
convince us that these categories of jobs that can and are likely to
be replaced are much wider, inclusive, as well as surprising than we
might have believed.

One way to look at this problem is to see that innovation, knowledge
acquisition and transfer, and the creation of new processes and
products are important in increasing productivity and living
standards, but that those same changes can also cause suffering
among those who cannot adapt or do not fit the needs of the new
economy and production processes.

Machines and automation of other kinds are replacing labor.
Investment in automation and labor replacement is taking the place
of paying wages.  That will affect the labor market.  That will
affect the ability and willingness of consumers to spend.

As machine become smarter and as they are driven more by computers
and software, they will also adapt to and learn new work more
quickly and easily.  The elimination of jobs by automation is just
as likely to speed up as slow down.

Automation is becoming less sector specific.  Whereas formerly,
workers, when laid off, might hope to train for work in another
sector, now those other sectors have been or are in the process of
being automated and employment there hollowed out, too.  One
consequence is that a prospective worker who is trying to follow the
advice that she or he should train and acquire education for her or
his next career, must hope to be lucky enough not to train for a
sector that will be automated by the time that training is acquired.

Many automated jobs will require less skill and knowledge that the
jobs that are replaced.  That has consequences: Pay will be lower.
Often the work will be part-time and without benefits, since the
tasks require less training and skills and because the workers are,
therefore, easier to replace.  As over-qualified workers fill these
jobs, former workers will be pushed further down to lower skill
employment tiers, where they usually earn lower pay.

Currently, the kind of artificial intelligence that is being used to
automate many jobs is special purpose, narrow artificial
intelligence.  But, as artificial intelligence becomes more
sophisticated, it will become more general and, for example, will
become better at learning, and these learning algorithms will make
it easier to automate a wider variety of tasks.

Some jobs are shifting toward a more "winner take all" income
distribution, where higher skilled workers receive a
disproportionate amount of the total compensation.  When you combine
that with the elimination of many jobs through automation and the
push of remaining jobs to lower skill and lower pay levels, you will
be looking at a very bleak employment landscape both for new workers
and for out of work workers.

Automation and de-skilling of the remaining jobs is leading our
economy toward more and more inequality.  There will be fewer jobs
that provide a living wage and there will be a small number of
positions that have very high compensation.  The income and wealth
distribution will no longer be a bell curve (a normal distribution)
with a fat middle.

Ford emphasizes that it is not just the routine, repetitive, low
skill jobs that are being automated.  Machines are increasingly able
to replace work that we believed was cognitive and required
intelligence.

The commonly proposed solution for these problems is more education
and training.  But, fewer jobs will be available even for those who
have more education and training.  Furthermore, many of us are not
capable of acquiring and making use of that education and training.
That is a solution that may help some individuals, but it does not
solve this kind of problem for our society.

And, apparently, in all advanced societies there is a race of sorts
going on between technology and education.  As technology advances
and automates jobs, education (and those who hope to find employment
by gaining that education) must adapt by finding yet other jobs that
have not been automated and for which it can offer training, skills,
and learning.  This race is augmented when educational institutions
help train those who will develop the technology that automates and
eliminates jobs.

And, Ford believes that we may be nearing the point where computers
and the software that runs them will become capable enough so that
it can effectively train itself to eliminate the next task or job.
I suspect that's a scary thought only for the future, but I'm not
too sure; perhaps it's beginning to happen already.

Perhaps there is a set of jobs that cannot be automated, but we do
not yet know which jobs, and one of Ford's points is that we cannot
predict which jobs those are.  He is also claiming that it is likely
to be an increasingly smaller set of jobs that provides meaningful
work and a living wage for fewer and fewer people.

We are moving toward a more economically unequal society.  That
means that some (the wealthy) will have the economic power to both
maintain and increase their share of the income distribution and
wealth.  Their increased economic power will give them more
political power, which they will use to ensure and further increase
their economic power.  That's called a vicious circle.

Another worry is that a successful economy requires consumers who
are capable of and willing to purchase the goods it produces.
Either that, or you've got to be able to export.  Companies must
sell their products in order to succeed.  And, when only a small
class of wealthy individuals can afford to buy those goods and
services, successfully selling them will become increasingly
difficult.

Exacerbating this is the fact that the share of compensation that
goes to labor, as opposed to owners of capital and executives, seems
to be dwindling.  And, that shift seems to be driven, or at least
enabled by technology.

There will be, as always, cycles and ups and down periods of the
economy.  But, there are indications that the periods of recovery,
especially with respect to job creation, will become shorter and
less robust.  Ford refers to studies that show that higher
inequality is correlated with shorter periods of economic growth.
That's something that should worry those who are just entering the
work force as well as those with children who will do so in the
future.

And, lurking in the background is the thought that we need more and
more economic activity and more productivity in order to provide the
increased number of jobs, or perhaps just to maintain the same
number of jobs.  But, if a perma-growth economy is the only solution
and the only "successful" future that we can imagine, then we are
headed toward a future in which we have more pollution, more CO2
emissions, more global warming, and a number of other unpleasant
consequences.  At some point we may have to accept the thought that
we will need to make do with less, especially less production and
consumption of goods, and also that we will need to share the wealth
more equitably.  For a sardonic view of this, see "The Collapse of
Western Civilization: A View from the Future" by Naomi Oreskes and
Erik M.  Conway.

If, and that is a very significant "if", there are political policy
choices that can be made to stop or slow down the advance of
automation and the consequent loss of jobs, it will be up to an
angry populace to demand it.  And, as the "Occupy" movement seems to
have shown, there are not yet enough people who are angry enough to
force that movement.  For a discussion of that, take a look at "The
age of acquiescence", by Steve Fraser.

For a detailed analysis of our currently increasing state of
inequality, you can read "Inequality: what can be done?", by Anthony
B. Atkinson.  Atkinson has a list of policy proposals, most of which
seem quite reasonable, although my guess is that political reality
makes them unlikely.  The proposal that is most germane to our
discussion about automation and the resulting lose of jobs is the
proposal that says that policy-makers should encourage innovation in
a form that increases the employability of workers.  (See Atkinson,
chapter 4.) How policy-makers and government decision makers would
do this seems a bit fuzzy, but Atkinson seems confident that asking
policy-makers to think through decisions that result in encouraging
one kind of technology over another will be enough push technology
development and use in the right directions (i.e., in the directions
that support increased employment).  The alternative, Atkinson seems
to feel, is to just allow any kind of technological development, no
matter how destructive to our society.

Of course, it is possible that allowing technological advancement
and automation to continue, which we may not be able to retard
anyway, really is the best thing to do.  Perhaps more automation,
more technology, and more productivity is the less bad option, even
if the distribution of the rewards from those changes becomes
increasingly skewed toward fewer people.

However, most of us count on work for some kind of emotional reward
as well an economic compensation.  We need to feel that we
contribute something of value; we need to feel that we are helping
to support our families; and so on.  Even if we could solve the
problems related to providing each or at least most of us with the
financial ability to afford a reasonable lifestyle in an economy
with fewer and fewer descent paying jobs, those job-less non-workers
still will be left feeling empty.

If you are interested in these issues related to jobs and
automation, then you may want to also look at (1) "The glass cage",
by Nicholas Carr and (2) "The second machine age", by Erik
Brynjolfsson and Andrew McAffee.

08/18/2015

.. vim:ft=rst:fo+=a: