The Automation Problem (1/2)

This was originally intended to be a single article. While writing, I found that the content I wanted to cover exceeded any sort of reasonable article word count by a considerable margin. If you find this text interesting, you can find part two here.

Automation is a word you’ve probably heard a lot in recent months. It’s been heralded as both the greatest threat and the greatest opportunity for humankind. What does it mean, why is it important, and how will affect all of us?

Some History

The concept of automation is not new. Human beings have been creating simple automatons since antiquity. Since Karel Čapek coined the word ‘robot’ in the 1920s, our knowledge of the technology that may come to underpin robotic systems has increased with public awareness of robots as a means of easing human labour following a similar pattern. Science fiction has taken the idea to extremes, often fantasising about the nature of such machines and what part they may play in the future.

Until the last 50 years, the implementation of real-world robotics has been limited. It was not until the late 1940s when William Grey Walter designed and built roaming robots – nicknamed ‘tortoises’ – that were capable of autonomous decision-making, that the field of robotics eventually moved from science fiction and elaborate stage art to the world of science.

Since then the fields of decision theory, physical robotics, computer science and electronics have come to gradually merge into a single field of research. This research has come to show not only that humans may be equalled by their robotic counterparts, but that robots may actually perform a vast array of tasks far better than human beings. For several decades, robots have been used in the mass-production of vehicles, food and other goods.

Recent Developments

Up until recently, ‘robots’ have all possessed fundamental limitations that prevent them becoming more useful to humanity. This is beginning to change with the rise of machine learning, a branch of computer science that deals with the theory underpinning computational intelligence. Recent leaps forward in both physical and algorithmic computing power have demonstrated not only that a machine may be programmed to perform a rigid, human-defined set of tasks, but that they may be taught to reason and learn in a manner beyond the limits of human comprehension. Image recognition, natural language processing, voice recognition, statistical analysis and real-time control systems are all domains in which modern artificial intelligence has demonstrated either considerable aptitude or the potential for it.

The Economics

Since the mid 18th century, industrial capitalism has come to be the dominant form of production utilised by modern economies to produce economic value: in short, goods and services that people rely on to survive. It depends fundamentally upon the relationship between workers; those that have the ability to work and capitalists; those that own ‘capital’, the property rights over land, equipment, buildings, intellectual property, and vehicles required to produce economic value. Workers are employed by the capitalist to produce economic value and are provided with a wage in return. This wage is then used to buy back the commodities produced through their labour that they require to survive. Although this may sound somewhat Dickensian, this fundamental model of economic organisation has remained relatively unchanged up until the present day.

How Did We Get Here?

Since before the dawn of humanity, human beings have been almost unique amongst the animal world in our use of tools to ease the production of economic value. We’ve crafted spears to hunt food, built huts to provide shelter, and utilised wood to create fire. Each innovation has made our lives easier and has allowed us to reach the next rung of the technological ladder. The invention of the plough made mass farming a reality and allowed us to end our previously nomadic lifestyle. The invention of the wheel made long-distance transportation of goods feasible, allowing us to live in cities without the threat of mass starvation. The invention of the tractor allowed most of our population to concentrate their time on things other than food production.

At each stage, the effect of innovation has been to amplify and optimise existing human labour, further reducing the demand for human beings in each occupation. This has provided the conditions for an incredible diversity of human occupations to flourish, allows us to satisfy ourselves in innumerable ways. At each stage, technology has opened up more possibilities for workers and has improved the living standards of almost everybody to some degree.

From this standpoint, industrial capitalism is simply the logical evolution of human society. It provides a mechanism for us to take advantages of economies of scale, the notion that the marginal cost of producing a good falls as total production increases. As in previous hunter-gatherer societies, humans may still experience and consume goods and services through the wage mechanism and the market economy in the same manner that the daily hunt provided goods and services to a nomadic tribe.

Yes, It Works. Sort Of

I’m sure at this point that many will raise arguments about the effectiveness of capitalism, about inequality, and about whether or not it continues to be a suitable system of economic organisation. Those arguments are, for the purpose of this discussion, irrelevant despite holding weight under other circumstances. Under capitalism, most people are housed, clothed and fed. Economic value does propagate through the economy in some way or another. From a bare-bones perspective, it works as a means of economic organisation, just as feudalism worked as a means of economic organisation. Most people are content and the world keeps spinning.

This happy dream all ends when you introduce automation.

When The World Stops Spinning

The pervasive problem with all this talk of economic progress is really quite simple. We’re making the assumption that human labour will always continue to be useful to society.  You see, capitalism – despite appearances – is quite fragile. To persist successfully, it requires that the relationship between worker and capitalist remains balanced and relatively unchanged.

This relationship must do the following:

  • Maintain the flow of goods and services (commodities that have inherent value) around the economy. For this to occur, human labour is required to produce this value in the first place. These two kinds of economic activity are highlighted in green in the diagram below.
  • Maintain the flow of currency (commodities that have derived value and exist as a form of contract commodity) in the opposite direction. This is highlighted in blue in the diagram below.

Screenshot from 2018-02-14 04:27:59

This is, of course, a gross simplification. In reality, modern industrial capitalism is far more complicated, with employers acting as quasi-employees of other companies (contracting) and other such complicated relationships. For the purposes of this discussion, however, those subtleties are irrelevant.


For over a century it has been suggested that automation may one day make human labour redundant. Countless Science Fiction authors have imagined a future in which human beings, no longer required to work, are free to live in a post-scarcity utopia where food, shelter, entertainment, intellectual stimulation and medical care are all free at the point of use to those that desire it. It’s a persuasive dream, if initially unbelievable.

So far, technological advance – as previously discussed – has only had the effect of amplifying existing human labour. This can be seen as an increase in worker productivity, a measure of the economic value produced by one person over a given time period. This trend has largely held true for almost all of human history. In recent years, however, this trend has stalled or even reversed in most major industrial capitalist economies. If technology has the effect of improving the efficiency of existing human labour, why might this be?

Most news analysis tends to point towards typical culprits; inappropriate infrastructure, shrinking financial sectors, and low interest rates after the 2008 financial crash. These explanations are most likely true, but I believe another important factor is at play. As firms increasingly employ the use of automation technology, human labour becomes less necessary to perform existing tasks. Classical thinking would have us believe that workers, newly unemployed, simply move into the new jobs created by the rapid march of technological progress. Unfortunately, this argument simply doesn’t hold up: as this research shows, the proportion of the workforce employed in the ‘Digital Sector’, a region of the economy that should be rapidly growing as technology advances, has almost constant since 2009. New jobs are simply not being created in vast numbers by recent technological progress to account for the numbers also being lost.

How is it then that UK employment remains high? For anyone working on a zero-hours contract, the answer is obvious: the quality, hours and wage of employment have dropped considerably. The so-called ‘gig’ economy is growing at an unprecedented rate. A multitude of protectionist government policies have been introduced across the world’s major capitalist economies in order to combat low employment, and the definition of ’employment’ has become skewed beyond recognition. In the UK, the number of workers in part-time work has increased considerably since the 1990s, even after taking into account population growth. The evidence is irrevocable: automation is already making human labour redundant. Dissatisfaction over the low quality of work and the appeal of protectionist job policies was likely a contributing factor in the election of Donald Trump in the USA and the decision to leave the European Union in the UK.

So, What?

The impact of human labour becoming redundant is substantial. With human labour mostly redundant, employers no longer have an incentive to employ workers. Without employment, workers no longer receive a wage and can no longer pay for goods and services. Without the ability to purchase goods and services, those that become unemployed through automation will fall into extreme poverty, further increasingly inequality in modern economies. This is an inevitable result of human labour redundancy, and it is not within the bounds of the existing model of industrial capitalism to provide a solution.

There is market no equilibrium. There is no limit. This trend of increasing inequality and unemployment will increase without limit until the only prosperous citizens of our society are those that either have skills unique enough to ensure continued employment, or own the robots and AI that drive the automation.

This Is Bad

It is not without irony that in a time when the human race is most capable of producing cheap commodities with little effort that it becomes totally incapable of distributing them equitably. Without employment, the very structure of industrial capitalism unravels.

The mechanisms that underpin modern industrial capitalism and the traditional relationship between worker and capitalist are ill-equipped to deal with the challenges posed by the automation revolution. We face a binary decision: radically change how our economies are structured, or face the dire consequences.

In conclusion: Hold on to your hats. The next 25 years are going to be interesting.

In the next article, I shall discuss possible solutions to the automation problem and consider ways in which is may be mitigated in the short term.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s