The Future of Work (and Engineering) in an Automated World
By Nick Brunt
While dreaming up the future of humanity’s relationship with technology in the 1940s, influential science fiction writer Isaac Asimov imagined the famous “Three Laws of Robotics”, a comprehensive set of rules designed to prevent any possible conflict between man and his ever-advancing machines. The laws stated that, first, a robot could not allow a human being to come to harm; second, a robot must obey the orders given to it by human beings; and lastly, a robot must protect its own existence. The laws were to be applied in that specific order, with the second two superseded by the first if necessary to create a binding, uncompromising safeguard for humanity as technology would progress beyond our wildest dreams. However, Asimov might be intrigued (if not truly surprised) today to learn that his unbreakable laws are routinely broken, albeit not in the most explicit or obvious of ways. Nonetheless, in this rapidly changing twenty-first century economic era, robots, computers, and the like are harming people, and nearly always at the behest of other humans. More and more, robots are stealing from humans one of the most consequential things that can be taken from them in a capitalist society: their jobs.
It is easy to see that we are living in a world in constant motion on economic, social, and technological fronts. The rise of computing and universal access to information has, in the past few decades alone, redefined nearly every major industry and precipitated the greatest period of economic change since the Industrial Revolution. And this change is accelerating. The exponential expansion of computing power coupled with a rather linear increase in job complexity means that ever more jobs may be better and more cheaply done by a few lines of code than a living, breathing, 9-to-5-working human being. In the view of MIT Technology Review senior editor Antonio Regalado, any work that is “repetitive or fairly well structured is open to full or partial automation,” which means few jobs are truly safe. What does that mean for the future of engineering, and employment as a whole? What further, far-reaching upheavals lie ahead in an increasingly automated world?
Since the dawn of the Industrial Era about two hundred years ago, fear of the supposedly apocalyptic ‘robot job takeover’ has, to some degree or another, been ever-present in Western society. Among this feeling’s first and most fervent proponents were the Luddites, the suddenly unemployed textile workers who, in protest of the loss of their high-skilled, high-paying jobs, took to smashing up the power looms that replaced them across eighteenth-century Britain. But as slighted and betrayed as the Luddites believed themselves to be, ‘technological unemployment’ as it was termed by economist John Maynard Keynes, has never lasted. Job displacement due to bold new technologies has always followed by replacement: the jobs simply have moved to some new sector or industry that rose as a direct or indirect result of the original change.
Since the dawn of the Industrial Era about two hundred years ago, fear of the supposedly apocalyptic ‘robot job takeover’ has, to some degree or another, been ever-present in Western society.
The tale of the machine-despising textile workers soon became somewhat of a cautionary one in economic circles. The name Luddite evolved into a blanket term contemptuously applied to anyone opposed to technological advancement, or the introduction of new technologies in a particular industry. Further, the idea that steady technological advancement does not lead to higher overall long-term unemployment, only the changing the composition of jobs in an economy, was christened the ‘Luddite Fallacy’. For most of the past two centuries, economists have been virtually united in their belief in this idea. However, we may be reaching a crucial point in history when the so-called ‘Luddite Fallacy’ is no longer a fallacy at all, and permanent job displacement and unemployment due to technological process could be an inescapable reality in the not-too-distant future.
Derek Thompson, economics writer for The Atlantic, formulated at very potent analogy for this future possibility in his 2015 article A World without Work. Consider the horse: one of mankind’s most valuable economic tools since the dawn of civilization – until it wasn’t. Millennia of technological progress only made the horse continually more valuable, as new plows and agricultural techniques only made the animal more and more productive. But that mutually beneficial relationship between animal and technology came to an abrupt end the day humans realized they didn’t have to keep making a horse increasingly efficient. They could just make a better horse, i.e. a tractor. We, humanity, could now be in the place the horse: on the verge of obsoleteness, and staring in the face of a revolutionary replacement we cannot seek to match: the automaton.
Permanent job displacement and unemployment due to technological process could be an inescapable reality in the not-too-distant future.
In their generously titled 2011 book Race Against the Machine: How the Digital Revolution Is Accelerating Innovation, Driving Productivity, and Irreversibly Transforming Employment and the Economy, MIT Business School researchers Erik Brynjolfsson and Andrew McAfee examined the American economic situation at either end of the Great Recession of 2008-2009. They found the striking discrepancy between declining employment numbers amid rising economic output could partially be attributed to – you guessed it – computer technology. As a result of these and other findings, there is growing doubt amongst economists everywhere that jobs will continue to be created at the rate they are swept into the dustbin of history by the newest computer or product or processor. Instead, explains equine analogist Thompson, “What may be looming is something different: an era of technological unemployment, in which computer scientists and software engineers essentially invent us out of work, and the total number of jobs declines steadily and permanently.”
So while engineering jobs may be safe for the moment, the rest of the working world might not be as lucky. In 2013, Oxford researchers published a first-of-its-kind study on the susceptibility of over 700 jobs to eventual automation (or to use their terminology, computerization). The conventional engineering disciplines – civil, mechanical, chemical, software, electrical and (evidently) computer – were all found to be among the safest 10% of all surveyed jobs, the least likely to be computerized. On the other hand, the outlook elsewhere was not remotely as bright. Alarmingly, even using conservative estimates, the researchers also concluded that a staggering 47% of U.S. employment was at risk from automation. While an economic shift of that magnitude may not happen overnight, the mere possibility of something similar occurring within several decades is unprecedented – and will require equally unprecedented social and economic reforms to ensure no one is left behind.
Even using conservative estimates, the Oxford researchers behind a landmark study on vulnerability of jobs to computerization concluded that a staggering 47% of U.S. employment was at risk from automation.
To underline the possible impact of such a change, let’s turn back the clock twenty years to innocent optimism of the early 1990s. In 1994, three countries, Mexico, Canada, and the U.S., signed the ambitious North American Free Trade Agreement (NAFTA). The trade deal was designed to make the signing countries more economically competitive on the world stage, and in at least that respect, it was a marked success. Nevertheless, this success was not without its drawbacks, with the United States in particular experiencing tremendous job losses in a manufacturing sector which had become uncompetitive with nothing more than the flick of Bill Clinton’s pen.
While NAFTA, as economists of the time predicted, worked as intended to increase the overall economic output of the signatories, for that very specific set of the American working class it was an unmitigated disaster. Suddenly, some people found that they had not only lost their jobs, they were no longer employable anywhere in the country. This impact was certainly not nationwide, but incredibly localized: while the economy as a whole grew, individual towns and regions bore the brunt of widespread unemployment and long-lasting economic depression. As of 2010, the Economic Policy Institute estimated that the United States had lost 682,900 jobs in the aftermath of NAFTA, mostly due to cheap labour south of the border, and some congressional districts had individually lost several thousand of those.
The regional collapses of manufacturing in the United States post-NAFTA presents a worst-case preview of what could happen with automation, albeit on a considerably smaller scale than what could yet occur. Furthermore, not only would the lost jobs be gone from the workers’ countries, they would be gone from the face of the Earth entirely. Consequently, NAFTA offers a prime example of how countries should not go about future economic transformation. Since automation is labour saving, it disproportionately targets middle- to low-income jobs, pushing them further down to lower income brackets. This could drastically increase income inequality. And unlike with NAFTA, future rising tides cannot be expected to raise all ships. Even if macroscopic economic changes have net positive benefits (such as in the case of trade deals, for instance), safeguards still need to be put in place on the governmental level to ensure the highly vulnerable middle class workers are not stranded with skills and experience that are no longer of any value, anywhere. This would mean the wide expansion of social security and unemployment benefits, which may tempt some future Western governments to implement some sort of Universal Basic Income system in the near future. In any case, public policy would seem to be the key to maximizing the smoothness of the inevitable transition towards computerization.
While an economic shift of this magnitude may not happen overnight, the mere possibility of something similar occurring within several decades is unprecedented – and will require equally unprecedented social and economic reforms to ensure no one is left behind.
In this future, how would people cope with perennial unemployment? Work itself is not just a means to the end of economic output or to the earning of money, but a daily purpose and routine that motivates people to contribute to society in other ways. When jobs leave, communities don’t just lose their economic vitality, they lose much more: pride, culture, a sense of togetherness, and faith in public institutions. Perhaps, then, what would be needed going forward are changes to societal attitudes towards work and employment. Another casualty of automation might be the near-religious devotion to work found in the West and the United States in particular, as people struggle with the societal shame of perpetual unemployment and unproductivity. Could automation be the end of the work ethic and spirit of capitalism described by Max Weber a century ago?
Only time will tell. The worst-case, NAFTA scenario of instantaneous collapse on an apocalyptic scale seems unlikely, simply because even the pace of technological development isn’t quite that fast or easy. Erik Brynjolfsson, co-author of the aforementioned Race Against the Machine, imagines the best-case scenario as somewhat of what he terms a “digital Athens”: mass participation in leisure, civics, and the arts enabled by a working underclass of slaves, only unlike the slaves of flesh and blood in Ancient Greece, we’d instead have slaves of microprocessors and metal. This is the idyllic future championed by ‘post-workists’, those who advocate and welcome the acceleration of automation as a means to achieve free time and freedom for all. The future of work, for all we know, will likely fall somewhere in the middle: not quite the listless dystopian landscape of empty office buildings, not quite the leisurely utopia devoid of woe and want. Which eventuality the world will tend towards will largely be a consequence of how well humanity can adapt to the radical economic upheavals that wait in the coming decades.
There is reason to be optimistic about the future. Cost factors will always be a severe limitation to global computerization, so if human labour remains cheap then human labour will always be the preferred choice for corporations. And, for the time being, there are some things that even the smartest machines simply can’t do. In 2014, the cofounder and executive chairman of LinkedIn, Reid Hoffman, was visiting a Huawei plant in Shenzhen China – and surprised to find not a line of robots, but about a 60-40 machine-human split on the factory floor. A full forty percent of the work was still being done by humans in high-skilled, high paying jobs, and he understandably wondered why. The answer was straightforward enough: “In the future, adaptability is key, and people are more adaptable.”
The future of work, for all we know, will likely fall somewhere in the middle of the two theorized extremes: not quite the listless dystopian landscape of empty office buildings, not quite the leisurely utopia devoid of woe and want.
Perhaps that’s the ultimate ideal of automation and productivity. Not just a future of only man or only machine, but a fine balance between the two. As long as that is the case, there will always be work for engineers and the rest whose unique skills can’t easily be replaced by some blue-collar, incapable-of-error HAL 9000 clone. Think creativity, leadership, innovation, and more. And for maximum job security in an automated future, the simplest advice might ultimately just be a rephrasing of the title of a presentation given by IT expert Craig Roth in January of 2017: if you don’t want to work for robots, “then make robots work for you”.
Now that’s a future Asimov would be proud of.