Thomas Klikauer and Meg Young: Automation and Looming Unemployment
(Dunk)
March 19, 2022
By Thomas Klikauer and Meg Young
Perhaps even before Nedd Ludd and his Luddites destroyed machines, factory owners have introduced the use of a new machinery – or what we now call as an automation – to lower wages and fire workers. Luddites were never against the machines in-itself (as Kant would say). Instead, they were against capital’s use of the machines as a deliberate strategy and, at times, as an excuse to fire workers. Ever since, employment and automation are in a tension-laden relationship.
Today, dystopian horrors spins that automation. They will lay waste on the global economy, replace the manufacturing, kill many jobs, create mass unemployment, and further atomize the so-called labor market – a market in which nothing is for sale. Today and in most cases, workers can no longer be purchased. Well, apart from the forty million slaves that labor every day. Beyond all that, it is estimated that the impending automation will lead to severe economical, social, human, and political consequences.
On the whole, nearly all labor-saving technological innovations which are based on automation can be divided between two groups: firstly, there are labor-substituting technologies. These sort of automations replace human labor. They abolish certain jobs entirely. Secondly, there is also an automation that is labor-supplementing or even enhancing. This a version of automation that is designed to make work and, more importantly, workers more productive. The second form of automation cranks up rather than destroys work.
Both versions will appear as a combined force in what became known as the 4th Industrial Revolution – also known as 4IR, or Industry 4.0. The 4th Industrial Revolution has been fancied by the likes of the World Economic Forum, McKinsey, the Boston Consulting Group, PriceWaterhouseCoopers, Deloitte, etc. Unsurprisingly, none of the 1st, 2nd, and 3rd revolutions were actual revolutions. There was no sudden change in political power as a result of a population revolting against an oppressive government or workers rebelling against an oppressive regime called capitalism. Yet, the idea of four revolutions lived on:
It all started with the 1st industrial revolution of the first machine age defined by a transition from hand production to machines driven by water and later by steam power.
The 2nd revolution (1871 to 1914) was defined by three key ingredients: a) Taylor’s strict division of labor; b) tasks fragmentation; and c) standardization under Taylorism and Fordism.
The 3rd revolution is known as the Digital Revolution. It occurred during the late 20th century.
Finally, there is today’s 4th revolution which moved industries towards automation and data exchange in manufacturing technologies based on processes which include cyber-physical systems, Internet of Things, cloud computing, cognitive and algorithmic computing, and artificial intelligence.
Particularly since the introduction of modern machinery, labor productivity climbed sharply from the year 1870 onwards – without any let up. Over this very long time, productivity even increased fifteen times in the United States and eighteen times in Europe.
Alone - this might indicate that automation is more utopian (positive) rather than dystopian (negative). The expectations of workerless and fully automated factories started to have appeared as early as the 1930s. At some point, this would not result in the economist John Maynard Keynes’ prediction that there would be a 15-hour-week. And that this would have happened by the year 2030 – a mere eight years from today.
Continuing its path towards automation, the 1950s and 1960s saw an anticipation of automated factories as General Motors was setting up the world’s first industrial robot in the year 1961. Twenty years on, and during the 1980s and 1990s, the structure of some work organizations began to be more and more defined by an extensive use of ICT.
Information and communications technology (ICT) is based on the integration of telecommunications, business software, and computer networks. Again, this has changed work. In the year 1962, technologically-advanced corporations like AT&T, Exxon, and General Motors employed hundreds of thousands of workers. Almost sixty years later, in 2022, Internet giants like Google and Facebook have workforces numbering in the tens of thousands. Worse, companies like WhatsApp and Instagram – that have been sold for billions of dollars – employ only a fraction of that.
Yet, all of this did not create mass-unemployment in OECD countries – the club of the forty richest economies. In one OECD country – the USA – the very opposite occurred between the year of 1989 and 2017: there was an increase of 118 million jobs in the United States. Yet, predictions made during the early 1990s suggested that warehouse labor, for example, would be reduced by 25% in the USA due to automation. Again, the very opposite occurred: it grew by 27% in the decade up to 2000.
After that, it has even grown by 83%. The Coronavirus pandemic has only increased this trend. All this occurred despite numerous recessions (e.g. the Global Financial Crisis of 2008, the Coronavirus pandemic, etc.) and the introduction of new automation technologies.
Through automation’s economic history – which has defined numerous rounds of technological developments – automation technologies have not replaced work and employment. Instead, it has created new forms of work. This had already continued throughout The Second Machine Age with its rapidly unfolding developments of digital and robotic technologies. Today, this include machine learning, computer coding, programming, artificial intelligence (AI), sensors, cloud computing, nanotechnology, and, of course, the Internet of Things.
Yet, it might be dangerous to read from past appearances of automation as to what will happen in the future. There simply is no deterministic inevitability about how automation will influence future employment. One of the main issues relates to AI and how it is applied in automating production.
Unlike in the past when the introduction of new machines were deployed in factories, this time around it is different. AI-guided robots are now capable of reproducing non-routine physical, cognitive, and even emotional labor. This is a qualitative difference unseen in the history of automation.
These smart machines rely on a combination of software, sensors, and robotics. They can emulate complex actions performed by humans. We see this already in self-driving cars that are capable of perceiving and responding to a wide range of highly different live traffic situations. Yet, it gets even more interesting. This sort of AI-driven automation is capable of machine learning. This is the key. AI-automation develops and betters itself autonomously.
Of course, nearly all of this is linked to economic factors. There is an ever increasing affordability of robots furnished with the ability to perform ever more complex work tasks. Gradually, these machines will become cheaper as usage increases. Secondly, the previously rather extensive pre-programming of these machines will become more affordable. This will make them more attractive to company management.
This may well automate away jobs to an extent not seen during the previous three machine ages. It is not surprising that McKinsey, projected that about half of all jobs in the US – mostly routine and semi-routine, manual, and cognitive jobs – are vulnerable to be replaced by robots. McKinsey also suggests that a Tayloristic task-based jobs will be replaced by automation.
Yet, there are also un-automatable tasks. These will, to a large degree, complement future automation. The Roosevelt Institute, for example, suggests that automation will not result in a gigantic labor displacement leading to mass unemployment. Of course, there is no determinism in all this. Automation is not inevitable, it is not a natural process, neither is it a God given. More importantly, automation and its introduction into companies and corporations never occurs independent of capitalism, corporate management, and managerial corporate cost-benefit thinking.
In short, the introduction of AI and self-improving robotics is far from inevitable. It will eventually be determined by their value and cost to capital – as it has been done since the invention of capitalism. Even today, the utilization of AI-driven industrial robots is still extremely costly as they incur very substantial programming needs.
As a consequence, many companies and corporations may shy away from costly and, at times, rather maladroit and clumsy robots. Some people are still waiting to see a robot that can simply open a door or read a book. Together with the serious costs of such an investment, all of this makes the exploitation of an existing workforce by capital, as a far safer bet than heavy investment in some sort of future technology.
Of course, ever since capitalism came into being, corporate management has always been keen to replace militant workers and their trade unions by introducing labor-saving machinery. In short, when workers are militant, better organized, and sufficiently unionized, employers have always been somewhat more devoted on labor substituting automation.
In general, the introduction of workplace technology – including AI and automation – has never just been the outcome of dazzling engineering feasts. Instead, new machinery, automation, and even AI is shaped by the imperatives of capitalism. In any case, corporate capitalism still depends on its ability to reproduce itself through globally expanded productivity regimes shaped by its eternal quest for profit maximization.
Interestingly, pressure from corporate shareholders can even act as a significant blockade towards investments in new automation. Between the years 2006 and 2016, for example, there has been a noticeable drop in corporate investments in information processing technologies, as well as into industrial equipment (robots, etc.), as a proportion of all new investments made in the USA.
Yet, The USA’s rather small productivity gains since the 1980s have been achieved largely through two aspects: firstly, there has been work intensification; and secondly, it has been achieved through a managerially-enforced one-sided and top-down flexibility imposed on a given workforce. In some cases this has been linked to regimes of flexible despotism. In other words, moderate productivity gains did not come from technical investment into robotics, automation, and AI. Worse, during the first decade of the new millennium, productivity growth even declined and technological innovations drove global growth by precisely minus 0.2%! – in short, no growth at all.
At roughly the same time, those countries that have invested heavily into industrial robotics – Germany, Japan, South Korea, etc. – were also those countries with the greatest trade surpluses, the greatest job security, and the best quality of working life in the manufacturing sector. Rather un-surprisingly, these three countries aren’t prime examples of the ideology of neoliberalism.
Following closely, China is investing heavily in fields like AI and robotics. It is targeting to become a world leader in the application of robot technologies. Since the year 2013, China has been the largest market for industrial robots. Chinese manufacturers have been buying up to a 30% share of the total supply of industrial robots in 2016. Its robot density – the number of robots per 10,000 employees – sky-rocketed by more than 270% from 2013 to 2016. And, this is set to continue.
Elsewhere, the marked downturn in the manufacturing workforce of many OECD countries shows the impact of wider political and economic pressures rather than a craving for the introduction of new technologies. Since decades, the share of workers employed in manufacturing has fallen across the high-income world. In the USA, for example, it has fallen from 22% in 1970 to 8% in 2017. In the UK, it fell from 30% to a meager 8%.
While the end of work and the apocalyptic prediction by many of the so-called experts about stratospheric job losses do not seem to be on the menu, there is however a renewed interest in the utopia of automation. There might even be some dreams that the dynamics of automation released by capitalism can create potentials to progressively liberate society from the yoke of capitalist work. Yet, under capitalism, the dreams of a Fully Automated Luxury Communism can never be realized. To achieve this, something entirely different is needed.
In the end, the sturdiness of capitalism indicates that any constitution of a political hegemony from the progressive side cannot be separated from grass-root organizing – organizing from below. This may rely on what German philosopher Ernst Bloch once called as Utopian Imagination. In other words, automation will neither bring mass unemployment nor will it lead to a Fully Automated Luxury Communism.
Follow BuzzFlash on @twitter
Continue the conversation at the BuzzFlash Nation group on Facebook