‘Robots Are Coming For Your Job!’ – Liz Ross

What is the future of work? What will happen to the jobs we do today? Is the future going to be a technological Garden of Eden as Marx sketched out in the 1800s or a dystopian, violent hell-hole? Three different writers have given their opinions. In 1931 John Maynard Keynes predicted a future of increased leisure time because new technology would take over much of the work, solving our ‘economic problem’ or the need to work. Sixty years later Jeremy Rivkin was far more pessimistic, pointing to a ‘fast polarising world dividing into two irreconcilable and potentially warring forces – a technological elite in control facing a growing number of permanently displaced workers who have little hope and even fewer prospects for meaningful employment in the new high-tech global economy.’

Twenty years on, Paul Mason proposes in his book, Postcapitalism, that the internet, automation, robots and Artificial Intelligence are creating a new economy that cannot be controlled by capitalism. New forces are at work replacing the old class struggle between capital and the working class with a network of communities. Mason suggests this will, unstoppably, lead to a post-capitalist world. He adds that the regular and recurrent crises under capitalism would dissipate towards a high productivity, low working day as capitalism itself ‘withers away’.

These issues and questions – or variations of them – have constantly greeted every new technological advance. More than in any other system, under capitalism new technology has both made life easier and opened up unparalleled productive forces, while at the same time, as Marx points out, is ‘used partly to throw the worker onto the streets…partly to break down his special skill…partly to subject him to the thoroughly organised despotism of the factory system and the military discipline of capitalism.’ Capitalism thus converts the development of the conditions of labour, what should be a process that adds richness and freedom to the producers into ‘an alien circumstance to the workers…’

In contradiction to the notions of Paul Mason, Marx, in his discussions on the trajectory of capitalism towards the increasing replacement of workers with machines, pointed to the existential crisis of the system that would result if machines did totally replace human labour. That is, that labour is the only source of profit under capitalism. A totally robotised or machine-operated system would produce no profit throwing capitalism into crisis and collapse. As Michael Roberts, a Marxist economist whose work I highly recommend, comments, robots do not do away with the contradictions within capitalist accumulation, the crises would begin well before robots took over. ‘So an economy increasingly dominated by the internet of things and robots under capitalism will mean more intense crises and greater inequality rather than super-abundance and prosperity.’

The alternative to such barbarism is, of course, socialism. And it was Marx who argued that there is an alternative – and much better way – to enjoy the fruits of new technology and in the process create a whole new system. As he said, capitalism creates its own gravediggers in the working class and a constant throughout capitalism’s history has been the heroic struggles of this class against the onslaughts of the ruling class. From the Luddites who smashed the machines of the textile factories in the early 1800s to the potential of the logistics and new tech workers of today – like Polar Fresh, Amazon workers in Germany and Verizon in the US, we have seen workers rise up time and again to challenge capitalism at its very heart. And bring with them the promise of a better world.


Technological innovation is nothing new to human societies. Developing from the first use of tools to fashion the world around us for our own use – what we now call work, there has been a history of revolutions in the organisation of these processes of work. We are now, some argue, with the arrival of robots and Artificial Intelligence – and possibly quantum computing, approaching our Fourth Industrial Revolution.

The question is, will this revolution be any different to earlier transformations, especially in its promise of a better life. Harry Braverman in Labor and Monopoly Capitalism explains how, to date, new developments in tools and machinery have largely been used for the benefit of the ruling class. He writes, ‘The capacity of humans to control the labour process through machinery is seized upon by management from the beginning of capitalism as the prime means whereby production may be controlled, not by the direct producer but by the owners and representatives of capital. …thus divesting the mass of workers of their control over their own labour.’

And the early developers of the computer were part of that process. Computers – or at least their forerunners – have been around a lot longer than the large-scale introduction from about the 1970s. About 150 years ago saw the development by Charles Babbage and Ada Lovelace (programmer) of the world’s first analytic and difference machines, the forerunners of the computer. Today we are facing a world where artificial intelligence devices, including robots, are predicted to reach parity with humans in 25 years time, do most of the work in 35 years time and outnumber human workers 10:1 by 2070, eventually to completely replace us in the workplace.

At the same time as he developed his machines Babbage – amongst others – wrote extensively about the organisation of work in the new machine-dominated factories. Without re-organising work involving a detailed analysis of every part of the work process, then assigning each worker to just one or two tasks, with the decisions and ultimate control of the work process in the hands of the employer – without all this, the efficiencies and financial benefits of factory production would have been lost to the ruling class.

In his book, On the Economy of Machinery and Manufactures (1832), Babbage illustrates one such process – what’s now known as the Babbage Principle. As he explained, skilled workers spend part of their time performing tasks that are below their skill level. He proposed assigning the lower skilled tasks to less skilled workers (on lower pay of course). Say in a workplace of 5 skilled workers you could have one skilled and four semi- or unskilled workers, you could do the job more efficiently and reduce your labour costs. His proposals, like those of Frederick Taylor of ‘scientific management’ fame, focussed on increasing control of the labour process by the employers to extract maximum profit – and of course to the disadvantage of workers.

Karl Marx was one of those ‘other writers’ at the time of Babbage. But his insights into the workings of capitalism, the use of machinery, lead, not to more and more management control, but to how we can build on the strategic strength of the working class, for a world where we control the fruits of our labour, including of course the use of robots and all the other new technologies, in the interests of all and not just profits for the rich.

Although these writers were there at the beginning of the first Industrial Revolution, much of what they – and particularly Marx – say is still relevant today as we face a Fourth Industrial Revolution with its questions about jobs – will robots take them all, how many jobs will disappear, what will be the quality of the new jobs, will the gains be more evenly spread – or will there be even more have-nots? How will work itself be transformed?

What about working class power? Will unions be irrelevant, will work be so different that, as Paul Mason argues, there’ll be networks of communities, not even a working class – or a ruling class?

To begin to answer some of these questions, we need to look at some of Marx’s analysis of capitalism and machinery. On the one hand, Karl Marx recognised, the technological productivity of capitalism is one of its most central virtues. Without the surplus capitalism created, egalitarianism simply means the generalization of want.  The Communist Manifesto explains that ‘the bourgeoisie, during its rule of scarce one hundred years, has created more massive and more colossal productive forces than have all preceding generations together.’

But then he also pointed out that:

‘Modern industry never looks upon and treats the existing form of a process as final…[It] revolutionizes the division of labour within the society, and incessantly launches masses of capital and of workers from one branch of production to another…[This results in] incessant human sacrifices from among the working class, in the most reckless squandering of labour-power and in the devastation caused by a social anarchy which turns every economic progress into a social calamity.’

In other words – capitalism both creates and destroys in a totally unplanned and chaotic way, destroying the lives of many both through joblessness and the creation of millions of mindless, soul-destroying jobs at the same time as opening up massive new possibilities for new and better jobs, for a better life, for a life of plenty. And each new technological development brings these contradictory outcomes to the fore.

So, let’s start with some of the positive potential of the advent of robots and artificial intelligence.  One short list would include the ability of robots and other machinery to do most, if not all, ‘dirty’ or dangerous jobs. Mining, deep sea diving, space exploration, cleaning, chemicals industry, removal of asbestos, land mines and the like. A whole lot of construction – as highlighted by last week’s tragic deaths of 4 building workers. Or the road toll which could be ended with driverless vehicles.

Michael Roberts describes at least four areas of massive change already introduced and rapidly expanding.

Industrial robotics and 3D printing which can make high precision, endlessly customisable products – and provide a whole new area of employment.  The advances in medical prostheses through 3D printing are just incredible.

Baxter – general purpose robot – has ‘eyes’ – LCD sensors. You don’t need to program Baxter – the human worker moves the robot’s arms showing it which moves do the job (picking up/putting down objects). Baxter memorises these patterns and can communicate that it understands its new instructions. Still just simple movements but you can add attachments to arms that increase range of movement. It’s currently being introduced to Haigh’s Chocolates in Adelaide.

Roberts goes on to what’s called the ‘Internet of Things’ which offers the possibility to connect machines and equipment to each other and to common networks, allowing for manufacturing facilities to be fully monitored and operated remotely. That’s just one possibility from the Internet of Things.

Adding this networking capability to the advances in data storage and manipulation creates what is known as ‘big data’ – the collection and analysis of large datasets. In health care – and many other sectors – big data is changing R&D, clinical care and forecasting; in some cases leading to highly personalised treatments and medicines.

And in infrastructure, the creation of Intelligent Transportation Systems, could see the introduction of smart grids, which could help save on power infrastructure costs and reduce the likelihood of costly outages; and efficient demand management, which could dramatically lower per-capita energy use. And driverless transport – cars, trucks.

And the list goes on. Almost every area of production – and many we haven’t even thought of – could gain from the introduction of robots and Artificial Intelligence devices. Some even predict the arrival of super-intelligence with quantum computing, an explosion of intelligence far outreaching human ability and the prospect of ultra-intelligent machines which would be, as one writer suggest, ‘the last invention that man [sic] need ever make’. [IE., machines would be able to do everything and more than humans can do.] [New chip from Microsoft – Field Programmable Gate Arrays – FPGA – can translate five bn words into another language in 1/10th second. Can confer 10x AI capability of largest existing supercomputer.]

The video that follows is a fairly simple example of one of the many areas of work that, it’s predicted, will never need to be done by people again.


Automation can lead to completely new – and skilled jobs. Computer typesetting, for example, has put a lot of printers out of work. But we now have many new jobs enabled by this technology, like web designers and e-book publishers. The introduction of ATMs which many thought would replace bank tellers has actually led to a greater number of ATMs – and bank tellers. Tellers’ jobs have been redefined, now including a range of higher skilled tasks.

IT jobs have increased by 77 per cent between 1997 and 2012 – these aren’t jobs within the IT industry, but IT and developer jobs created in the companies outside the tech industry. After first being concentrated in manufacturing, many new IT jobs are appearing in white collar administrative and service industries.

New technology doesn’t always mean a reduction in skills level as the examples I’ve just given show. It can mean the creation of entirely new skills or a skilling-up of workers already employed. Early factory production lines did employ mostly unskilled workers, but electrification allowed greater automation and an increased demand for relatively skilled blue-collar production workers to operate the more complex machinery. And laid the basis for some of the growth in unions – certainly here in manufacturing.

This in turn meant that as factories grew in size and number, as transportation spread and improved, that managing more complex production and transport networks needed more white collar administrative, technical and managerial jobs. The demand lent weight to the push for higher levels of education – in the US the high school movement – and after that greater tertiary access and training.

While this is some of the upside of increased automation, what are the downsides? What about jobs? New technology, everyone accepts, does lead to the destruction of whole sectors of employment and can lead to serious dislocation for those losing their jobs. In Australia, for example, research based on closures in the manufacturing sector, shows that one third of workers will move into equivalent jobs, one third will move into lower skilled, lower paid, possibly less secure jobs and one third will remain unemployed or in retirement.

There’s much more uncertainty expressed about the number and quality of jobs that will be created. Will the creation of new jobs match destruction of the old? And if not what does society do about transitioning workers from the old economy to the new? What happens when job creation lags job destruction, when skills are not easily transferable, when a proportion of the new jobs could be precarious, lower-skilled with less pay and worse conditions? Uber, Airtasker, etc, etc.

The picture I’m going to paint is complex and contradictory. Overall transformations within capitalism have led to a higher standard of living, better jobs, shorter hours and the like, and most research backs up these kinds of results for a future automated society. But there are no certainties.

Let’s look at some of the facts. Are robots taking over all the jobs now? The use of robotics has doubled in the top capitalist economies in the last decade – and may well increase exponentially from now on. It’s still got a long way to go however. In 2015 in the US, robots had only penetrated ten per cent of the manufacturing sector – the Bank of America predicts an exponential increase to 45 per cent of manufacturing tasks by 2025.

Nonetheless the robots are coming, so what of the jobs that could disappear or completely change? One of the most influential studies of the impact of robots on jobs is that of Carl Benedikt Frey and Michael Osborne (2013). Their survey indicates that in the next two decades 47 per cent of employment is in the high risk category – that is the jobs – or large parts of those jobs, that could be replaced by robots. This covers 702 occupations and in the US the equivalent of 138m jobs. In Australia the estimate is around 4.6m jobs – out of a workforce of roughly 11m. (Startup AUS – CEO Alex McCauley)

Frey & Osborne conclude that, ‘As the comparative advantage of human labour in tasks involving mobility and dexterity – ie manufacturing – will diminish over time, the pace of labour substitution in service occupations is likely to increase even further.’

So it will be the middle layer jobs – work done by many white collar workers – clerks, lawyers, financial analysts, journalists, librarians or surgeons – which are becoming increasingly automated/computerised/robotised and facing disappearance.

It’s not just that robots are replacing people in certain jobs, it’s that the whole way of working is changing. W. Brian Arthur, a visiting researcher at the Xerox Research Centre says it’s far more subtle than the idea of robots and automation doing human jobs; it involves ‘digital processes talking to other digital processes and creating new processes,” enabling us to do many things with fewer people and making yet other human jobs obsolete.’

But, to date, is this all resulting in major job losses? In the period 1993-2007, Graetz and Michaels found that in 14 industries in 17 developed countries including Australia, industrial robots increase labour productivity, total factor productivity and wages. They had no significant effect on total hours worked but did impact the number of jobs. So in essence, robots did not reduce toil (hours of work) for those who had work, on the contrary. But they did lead to a loss of jobs for the unskilled and even those with some skills.  So more toil, not less hours; and more unemployment.

That picture of hours of work may be changing as 2016 data from Australia does indicate some fall in hours worked over the last twelve months. (Centre for Future Work) Henry Siu and Nir Jaimovich show that the jobs – those most sensitive to technological change – have collapsed in the US since 2001 leading to the jobless recoveries seen more frequently of late.

Siu and Jaimovich write: ‘We show that over the past 40 years, structural change within the labor market has revealed itself during downturns and recoveries. The arrival of robotics, computing, and information technology has allowed for a large-scale automation of routine tasks. This has meant that the elimination of middle-wage jobs during recessions has not been accompanied by the return of such jobs afterward. This is true of both blue-collar jobs, like those in production occupations, and white-collar jobs in office and administrative support occupations. Thus, the disappearance of job opportunities in routine occupations is leading to jobless recoveries…’

Nonetheless a more positive outlook on hours worked – they haven’t decreased – is supported in research by Jeff Borland from the University of Melbourne published just last month. Answering another proposition that this round of automation is advancing at such a rapid pace that we are facing a major disruption, he shows that the pace of change in the composition of employment and in job turnover is no quicker today than before the widespread use of computers. At least in trend figures.

More generally unemployment levels have stayed around 4-6 per cent in most developed countries. And most studies indicate that the majority of workers are still in permanent full-time and part-time employment, with casualised or precarious jobs staying roughly the same. Independent contractors, for example, have remained at about 10 per cent of Australian workforce, a figure which has been stable over past 10-15 years. (Andrew Stewart). Many of those taking up work at platforms like Uber, the so-called precarious jobs, are already working full time or permanent part-time or are retired. This is a pattern of work that has a long history.

So the answer for now is that while work is becoming increasingly automated, with the next wave of change coming to a range of white collar jobs, there is no mass job loss, nor are most people catapulted into precarious jobs with no future. While new technology eliminates jobs, it does not eliminate work itself – instead new jobs have been created as a result of technological change.

However in the 1950s many imagined we’d already be in a world of robots and artificial intelligence, a world of leisure and plenty. This clearly hasn’t happened and in fact we’re still on the verge of the next next technological breakthrough and many work longer hours, while wages growth has stagnated. What’s happened – well it’s capitalism, which rather than develop what people need, has chased profits, particularly in the defence industries where roughly 95 per cent of investment into automation in the US has gone.

Whether job loss and job creation does continue to balance each other out in the wider economy, depends on the health of the capitalist system – and many of the signs are not encouraging. The rate of profit continues to trend down and even a decline in wages growth has not boosted profit margins enough to encourage investment. Investment to create the new jobs, taking the US as an example, has dropped from around five per cent of GDP in 2000, to around 3.5 per cent in 2014. Similarly US productivity growth was the lowest in first quarter 2015 since measurements began in 1948, with over 25 per cent of the country’s industrial capacity lying idle. And this is typical of many other economies.

In other words – the real threat facing workers today – is not new technology, but capitalism itself. It’s the way that the ruling class organises and controls work, the decisions they make in the interests of profit that has – and will – determine whether or not robots and artificial intelligence more generally will destroy the lives of millions.

As Doug Henwood concludes, while there’s no evidence that info-tech is ‘hugely diminishing the amount of work needed’, ‘Sure wages and benefits stink, but that’s about politics and class power, not because of the latest generation of Intel chips or something fresh out of the latest TechCrunch Disrupt.’

It’s the actions of our side – the working class – that will have – and have had – to make the difference over the years when new technology has been introduced, that have garnered some of the benefits of new technology and now have the potential to do the same as robots and other artificial intelligence applications become increasingly common.

Two hundred years ago, in England with the Luddites, we saw the first nationally organised working class resistance over threats to wages working conditions and jobs from the introduction of new technology in the textile industry. Carrying on that tradition have been the industrial action, strikes, factory, mine and other workplace occupations of millions of workers worldwide and even some attempts at workers control – such as the BLF in the 1970s, where rank and file workers took over the organisation of work building the Opera House. Or rather like their Luddite forebears, in 1975 when printers at the Washington Post held their foreman hostage while they meticulously destroyed and burned the computerised press that had just made them obsolete.

The story more recently has not been as positive – we are facing challenges more than celebrating success. Attempts at unionisation in the heartlands of new technology has not seriously broken through the strident anti-union stance of the big IT firms. David Bacon, an IT worker and union organiser, vividly explains how this has happened.  [David Bacon, Truthout, 2 March 2011]

In the US, as the electronics industry began to grow in the 1950s in the heartland of new tech, Silicon Valley, inter-union battles drove out left wing activists and officials and fragmented the union movement. As a result, while the new high-tech industry was growing, the ability of electrical and electronics workers to organise unions in the expanding plants fell to its lowest point since the early 1930s.

By the 1970s there were again serious attempts to unionise amongst tech workers in the Valley, but, after winning some gains, many of the activists were fired and much of the production side of the industry was re-located to New Mexico. But unionisation attempts – and a fight for better wages and conditions – were revived in 1990, not all amongst production workers, but amongst workers such as janitors. In 1992 there was a bitter campaign, started when employers used workers’ immigration status to fire them in large numbers. The strikers also drew in the support of their local communities, ‘So a community coalition went to picket when our union couldn’t, supported the workers with a hunger strike and started a boycott of Apple products.’ The janitors won and remain unionised to this day.

In the same year Versatronex workers, again mostly immigrants from Mexico, Central America and the Philippines – were the first Silicon Valley plant where production employees went on strike and the first where a strike won recognition of their union. A year later the plant closed, offshoring to Asia. But the workers still saw the strike as a win – something our unions today could learn from.  ‘We said at the beginning that if the company was going to close, let them close’, said Sandra Gomez, a leader of the Versatronex strike. ‘But as long as the plant was open, we were going to fight for our rights.’

Bacon points to the importance of these battles: ‘Despite these attempts Silicon Valley is the fortress of the country’s most anti-union industry. Overcoming it has the same strategic importance that organizing the steel and auto industries in Pittsburgh and Detroit had in the great industrial union upsurge of the 1930s. For the working class and communities of colour in Silicon Valley to assert their own interests and to ensure that economic development meets their needs, the workers in the Valley’s plants must be organised.’

And the geeks do fight back as Tom Barnes tells in his thesis on IT workers in Australia.  ‘Angry geeks down mouses in industry first’. This was the headline in Workers Online, the New South Wales Trades and Labor Council’s website, on 9 April 1999. After nine months of negotiations through their union the ASU, with Global Systems management for an EBA with no result, workers banned the release of new software to the plant in Japan. When management ordered them to lift the bans, they, as the headline said ‘downed mouses’.

According to one worker, ‘What we managed to push them into doing was having a sort of Clayton’s EBA. It is a collective common law contract which everyone is covered by and we got put in provisions that any new staff that are hired have to be under the same conditions. Nothing can be changed without all staff agreeing to it…. It was really quite interesting that in nine months people went from zero consciousness to going on strike. It was the first strike of geeks in Australia definitely.’

But winning didn’t save their jobs. The company closed down its Sydney site and moved it all to Japan.

It may be more on the edges of the industry – in logistics, in the armies of support staff, workers like janitors and the like – where the breakthrough comes. Facebook’s bus drivers voted to unionise in 2014. They won an $18-20/hour rate which is above the minimum wage, but scandalously is still below a living wage for the region. And we know about Polar Fresh – hear about in the next session.

Then there is Deliveroo. Management proposed a change in courier pay structure – the existing £7/hour plus £1 per delivery was going to be reduced to £3.75 per delivery – riders took to the streets of London. They’d quickly worked out they’d have to work longer or faster to maintain a sustainable income.

New technology and the app-based economy played its part in the Deliveroo victory but the victory itself was based on a much more timeless concept: that the power of workers lies in their capacity to withdraw their labour. Some commentators have credited the gig economy’s growth to a trend towards worker individualism but the Deliveroo strike suggests collectivism still works.

Today the term Luddite is usually an insult implying backwardness or an anti-progress stance, though in fact this is a recent interpretation; only coming in in the 1950s alongside that wave of new technology. The real history of the Luddites tells a very different story. They weren’t just worker activists. The Luddites combined a politics of revolutionary insurrection, openly associating their cause with the revolts in France, America and Ireland They wrote their own manifestos arguing that supposedly neutral technological change was in fact political, shaped by the push by employers to assert greater management control over work.

The highpoint of the Luddite rebellion was 1811-12, with organised Luddism ending in 1813 – though sporadic outbreaks continued and a similar uprising occurred later in the South of England protesting against threshing machines – known as the Captain Swing Riots. All told 12,000 soldiers were sent in to put down the Luddite uprising, nearly 50 protesters were killed by government forces and after the show trials, 24 were hung, many imprisoned and 34 transported to Australia.

Nonetheless by the end of the uprisings thousands of frames, a significant proportion of the total number in England had been smashed. And although it is often argued that the Luddites failed, in Nottinghamshire many of the master hosiers [stocking makers] did not use the wide frames for some years and wage levels were considerably restored. The Chartist movement of 1839-1848 was undoubtedly a legacy of the Luddite uprisings.


The Luddite resistance raises many of the issues we face today – what role does technology play, why and how it’s introduced and what should the working class response be. While I’m not going to argue an anti-technology case, anymore than the Luddites did, we could do worse than organise nationally across whole industries and adopt as one of our demands, their call “For action against technology ‘hurtful to commonality’”.

In Australia the United Voice union at its national conference in 2015 said that ‘technology is a key driver in the changing work of our members. We need to do more than embrace it. We need to adopt an aggressive, future-oriented and proactive stance on technology. It’s not the robots we should be afraid of, it’s our inaction.’

[This article was originally presented by Liz Ross as a paper at Socialist Alternative’s Marxism 2016 conference.]