The Future of Work: Video Games in the Conference Room, Philosophy Lectures in the Machine Shop and Gourmet Meals in Underground Mines
For more than 100 years, economists, philosophers and scientists at all ends of the political spectrum have predicted a future where much of the work that has been historically been performed by humans will be taken on by machines and computers. In 1930, John Maynard Keynes imagined that his grandchildren would live in a world where the humans would struggle to find meaningful work to fill their time and the length of the work week would fall to 15 hours [i]. In 1950, mathematician Norbert Wiener, argued that machines would soon free society of ‘relentless and monotonous drudgery,’ allowing humans to spend our time on engaged in creative knowledge work [ii]. In 1971 anarchist Murray Bookchin suggested that industrialized countries were rapidly approaching a “materially abundant, almost workless era in which most of the means of life can be provided by machines,” in his proposal for a post-scarcity anarchism [iii].
More recently, in his 2015 book, “Rise of the Robots, Technology and the Threat of a Jobless Future,” Martin Ford acknowledges that the previous warnings of mass technological unemployment but argues that the current wave of technological improvement is different because, the exponential growth potential of information technology “is pushing us towards a tipping point that is poised to ultimately make the entire economy less labor-intensive” [iv].
Research on the potential for automation in current jobs seems to support Ford’s prediction. In 2013, Oxford University researchers Carl Frey and Michael Osborne analyzed tasks performed in 702 detailed occupations to estimate the probability of computerization, suggesting that about 47% of US employment is at risk of technological displacement [v]. More recently, writing for the consulting group, McKinsey & Company, Chui, Manyika, and Mehdi took a more granular look at the question, examining the percentage of time workers spend performing particular tasks and assessing the degree to which those tasks are susceptible to automation [vi].
This research is very useful because it allows us to begin to understand how technological advancement might impact work and the labor market more broadly. But while these studies tell us what work might be automated, they tell us very little about how jobs will be impacted. In reality automation is not likely to eliminate many jobs at all. Instead, as technological advances make work more efficient, most workers will just end up spending more and more of their time at work with nothing useful to do.
The idea of millions of workers sitting around workplaces with nothing to do for most of their day seems absurd and incredibly inefficient. If there’s not enough work to fill a person’s time why not combine jobs so workers can stay busy all day? It is certain that a good deal of that consolidation will happen, but there are legitimate and likely unavoidable ways that organizational friction will limit the continuations combination and elimination of jobs. Demands for continuous production and the need to keep workers around to handle edge cases, the real human limits on the concentration of emotional and cognitive labor, and security concerns demanding compartmentalization of work all create legitimate and likely unavoidable forms of organizational friction that will lead people to spend time either doing things that machines could do better or faster or sitting around work with nothing meaningful to do.
Suggesting that automation is unlikely to actually eliminate all that many jobs does not mean, however, that there are not interesting and important ways that automation can transform work and improve peoples’ lives. Once we acknowledge that automation is going to lead to people spending incredible amounts of time at work doing nothing all that useful (and I would suggest that a lot of this is already going on) we can begin to imagine ways that we can transform work and workplaces to make them less degrading, less alienating, and allow people to actualize their human potential while participating in wage labor.
Just Keep the Machines Running
In capital-intensive industrial workplaces like paper mills, oil refineries, and underground mines, the potential costs of lost production caused by allowing expensive equipment to sit idle for any period of time is incredibly high. Many manufacturing workplaces are actually already largely automated and production workers spending most of their time looking on, making minor adjustments, and handling edge cases.
As equipment in these workplaces becomes even more automated and equipment can run with little — if any — human intervention, because the costs of downtime are so high, we can expect to see employer to continue to keep a well-staffed workforce on hand, ready to make whatever adjustments, corrections or repairs are needed to keep the equipment running. Paying workers to sit around in case something to go wrong is simply cheaper than risking any avoidable downtime.
In modern paper mills, when everything is going well the whole production crew can watch from an air-conditioned control room (almost always equipped with a television or tablet streaming sports or a movie) as the machine runs along, only needing to emerge when there is a break or irregularity the paper or to start a new roll.
As paper-making technology advances, sensors improve, and the introduction of machine learning and AI enables paper machines to self-adjust to irregularities, we can imagine that workers will end up spending more and more time in those control rooms and less time on the floor. While employers may choose to reduce the workforce, the relative costs of paying workers to watch TV and be on the ready to address problems with the equipment are almost certainly going to be far less than experiencing any more downtime than is absolutely necessary.
With relatively little to actually do while automated equipment is running smoothly some workers have come up with interesting ways of filling their time. I have some friends who work in an underground nickel mine who found themselves with so little to actually do while automated equipment is operating properly that they were able to cook elaborate, gourmet (if incredibly unhealthy) meals using a variety of hot plates, crockpots, toaster ovens and other improvised pieces of cooking equipment. The types of dishes they can prepare are limited to ones that can be put on hold if they need to rush to address a production issue. But overall this group of workers seemed to find spending their downtime cooking elaborate meals fulfilling.
At one point this group of miners-turned-chefs came up with the idea to produce a YouTube cooking show to document and broadcast their industrial culinary creations. They even came up with at title, “Cooking on the Clock” and started patching together a pilot episode. Before moving ahead, however, they heeded the advice of their union representative who pointed out that, although their supervisors and basically everybody in the entire operation knew that the nature of the work required large amounts of down-time, broadcasting their antics might be taken poorly by senior management who were not as familiar with the production process. They began to worry that drawing the ire of senior management could result in directives to stop the cooking and spend their time polishing equipment or doing some other utterly useless, but time-consuming task so they abandoned plans for the cooking show but continued making gourmet meals.
In these capital-intensive industries, paying employees who spend very little of their working time doing any actual work in furtherance of the company’s mission is actually squarely within the employer’s narrow economic interest (whether they recognize it or not). This will likely continue unless or until technological advances in industrial equipment develop to the point that the equipment can handle all edge cases and maintain itself — a future that seems relatively impractical in the imaginable future.
The Human Limits of Emotional Labor
In some workplaces, automation can increase, not decrease the burden on employees. If we imagining AI and machine learning technologies being deployed widely enough in workplaces that humans are only called upon to perform the work that is so complex, so requiring of human judgment or so emotionally demanding that it cannot be performed by advanced computers, there are many, many fields where no human could withstand that level of work for a full eight-hour shift or a forty-hour work week.
Imagine a hospital that had been able to automate every aspect of performing surgery so there was practically no need for human surgeons. But that hospital recognized that the task of notifying families that their loved ones had died during surgery or telling a patient that had been admitted with a spinal injury that they would never walk again was something that should not be done by a robot, no matter how sympathetically programed. The person tasked with this heart wrenching work would need sufficient medical knowledge to explain to the family or the patient what had gone wrong, so this work could only be performed by a trained surgeon. The one surgeon at that hospital would spend all day, every day doing the one piece of work that the robots couldn’t, delivering devastating news to patients and families.
Of course, no human could withstand the emotional toll of performing the some of the most challenging emotional labor imaginable all day, every day. And while this dystopian robotic hospital staffed only by a single surgeon-turned-angel-of-death is certainly unlikely to materialize, concentration of emotional labor and increasing emotional fatigue is already a challenge facing many healthcare workers. This is particularly pronounced in the long-term care sector where healthcare providers are pushed to provide compassion and care to patients in the same way that they would care for their own parents. While this emotional connection can be healthy and generative in a real, presumably reciprocal, kinship relationship it can be daunting for healthcare workers who only give and do not receive emotional labor [vii].
Currently this intense emotional labor is only one portion of the job that long-term care workers perform. But technologies that either already exist or on the horizon could automate many of the other portions of the work, from checking vital signs to assisting patients in bathing to changing bed-pans. If long-term care workers see their job functions concentrated in the emotional labor-intensive aspects of the work, providing companionship and counseling to patients navigating the end stages of their lives, suffering from dementia and Alzheimer’s, and grappling with loneliness and isolation it is imaginable that many of these workers will begin to long for the days when their work was broken up by periodically changing bed pans.
While researchers examining job design have long argued that creating jobs offering employees rich and measurable work combined with meaningful feedback increase both employee satisfaction and employee productivity [viii], other observers have noted that there are limits to the extent which increasing the “richness” of a job can actually improve performance before completely overwhelming employees [ix]. Of particular interest here is the body of work looking at job design for workers in creative or professional fields. Elsbach and Hargadon even provide some evidence that scheduling “mindless work” into the workday can improve the productivity [x]. As increasing levels of automation eliminate busywork and less interesting task, we can imagine that the remaining workers will spend their days performing only the richest of work, likely with deleterious results.
Employers in the healthcare and other labor-intensive industries have long worked to push the boundaries of their employees’ physical and emotional limits. But as hard as they push, if workers’ efforts are concentrated in only the tasks that are too complex or uniquely human to automate, they will run up against the emotional and cognitive limits of the amount of work that their employees are able to safely perform. They will almost certainly find that limit well before the passing of a full work day or work week.
Other Legitimate (and Illegitimate) Sources of Inefficiency
As artificial intelligence and machine learning continue to transform the world of work, other forces will continue to create organizational friction leading to many people spending large amounts of their work days doing nothing particularly useful. In the financial sector and many advance research and security organizations, work is strictly compartmentalized at all but the highest levels because of security or privacy concerns or to avoid conflicts of interest. While many similar jobs in these sectors may be combined, this demand for compartmentalization places real limits on the possibility of job consolidation.
In other cases, cyclical or seasonal markets may lead employers to maintain significant levels of employment of workers with specific skills or in specific trades through downturns to ensure access to that labor when demand is strong. While firms generally prefer to lay off workers when demand falls, they also recognize that when the market turns around and demand rises again, everyone will be competing to hire the same population of workers. By keeping workers on the payroll through downturns, firms can be confident that they will be able to quickly scale up production in more attractive markets. This type of friction has long mitigated labor market volatility in the durable goods manufacturing, hospitality, and energy exploration and extraction industries.
In a much more nefarious scenario, employers engaged in particularly destructive enterprises are likely to retain high levels of employment even as automation reduces the need for human workers as a way to maintain their moral license as “job creators.” If it were not for their status as significant employers there is almost no chance that we would allow coal, oil and gas companies to continue to pillage the earth and pump millions of pounds of carbon into the atmosphere. Energy companies likely recognize that continued reductions to their workforce substantially erode the degree to which politicians and regulators are willing to tolerate their destructive behavior which can lead them retain more employees than they otherwise would be inclined to.
So, what are workers doing with their time?
To be certain, even taken together these factors will only limit, not eliminate technologically induced job reductions. But as evidenced by the nickel miners-turned chefs, tv-watching paper makers, and any number of other workers finding creative ways to fill time at work, the era of technologically induced idle time at work is already here. And as advances in artificial intelligence and machine learning further automate our workplaces, leaving behind only that are too complicated too irregular or too uniquely human for human attention, fewer and fewer workers will find themselves actively engaged in any sort of meaningful labor throughout their entire work days.
With more and more workers finding themselves with more and more time on their hands at work, both workers and employers have an opportunity to reimagine how work is organized and how people interact with their work. Taking advantage of this opportunity, workers and their employers could find ways to make work less degrading, less alienating, and allow people to actualize their human potential while participating in wage labor.
Unfortunately, however, in the vast majority of workplaces, moralist assumptions about the value of work and the perils of idleness combine with genuine fears of displacement through redundancy force workers to hide the amount time they have at their disposal and encourage employers to fill any of that downtime with meaningless busy work.
In the summer of 2013, anthropologist David Graeber published the essay, “On the phenomenon of bullshit jobs: a work rant,” in Strike! magazine, blowing the whistle on what is actually happening in workplaces around the world. Graeber argued that “productive jobs have, just as predicted, been largely automated away… But rather than allowing a massive reduction of working hours to free the world’s population to pursue their own projects, pleasures, visions, and ideas we have seen the ballooning of…’bullshit jobs’” [xi]. These ‘bullshit jobs’ Graeber argues are not inherently bad or unpleasant jobs, nor are they particularly nefarious jobs. Rather ‘bullshit jobs’ are jobs that even the people who perform them think are meaningless.
The article clearly struck a nerve. It was immediately a runaway success. The week it was published the million hits on the story crashed the Strike! magazine website multiple times. Within weeks it had been reprinted in newspapers all over the world and translated into more than a dozen languages. Emboldened by the overwhelming success of the original article, in 2018 Graeber published a book-length work, “Bullshit Jobs” complete with a taxonomy on the types of bullshit work pervasive in workplaces across the western world: “flunkies, goons, duct tapers, box tickers, and taskmasters” [xii].
As technological advances make work more and more efficient, but the frictional factors identified above limit just how many workers can actually be displaced more and more workers will find themselves with free time on the job. Unless we can break through the antiquated and moralistic attitudes about the virtue of work and the perils of idleness and have a real conversation about what we want our workplaces to look like, we’re likely to waste this opportunity and continue to fill this time with bullshit.
What could we do with all that time?
If we were able to get comfortable with the fact that labor saving technological advances combined with barriers to total automation are going to leave workers with a lot of extra time on their hands, we could begin to collectively reimagine our workplaces to take full advantage of the potential benefits of automation. By exploring ways to intermingle work, leisure and family time and space we can allow workers to truly enjoy their down time and attend to personal and family commitments while at work.
In some cases, the simplest solution for making healthy use of idle times may be for workers to just go home when the work is done. Healthcare workers who can only reasonable perform 4 or 5 hours of complex and intense physical, cognitive and emotional labor should probably just leave work when their shifts are done. In cyclical industries where slow periods mean workers have less work it might make sense for those workers to only come to work for a few hours a day a few days a week during slow periods.
But in workplaces where workers need to be on site, we could go a long way in making work less exhausting, alienating and degrading by encouraging workers to make use of their down time in ways that make sense for them. To the extent that extraneous activities do not materially disrupt the work process who cares if workers are playing video games in the conference room or reading philosophy in the machine shop?
Childcare can be a major cost for many young families and lack of reliable childcare is a major contributor to employee absenteeism in many industries. In workplaces with a critical mass of young parents, employers could go a long way in making workplaces more family-friendly by allocating an appropriate space for children and parents to spend time together during the day. Parents could take turns supervising the children while balancing their work responsibilities.
Mixing work and leisure space can also allow workers to more freely and openly share their skills and resources with each other. An employee who is particularly good at auto repair may offer to fix her co-workers’ cars during down time while another may offer haircuts to his co-workers.
This would not only be a more humane way to structure work, it could also yield financial benefits for employers. The role of flexible work arrangements on reducing turnover and improving job performance is well documented [xiii]. These ancillary activities could also create value for employers. What if rather than shunning underground miners for cooking decadent meals at work, their employer encouraged them to launch their, “Cooking on the Clock” YouTube show and negotiated to share a piece of advertising revenue?
One incredibly practical possibility would be to encourage and support employees in engaging in continuous education programs while at work. Many employers currently offer subscriptions to online courses or tuition reimbursement for employees, but in most cases, employees are required to participate in those courses on their own time or on a very limited basis at work. If workers and employees could come clean about the fact that they are only actually working a fraction of the day, this continuous education could be understood as a core, not ancillary part of the work and workers could take on more ambitious learning objectives.
In his 2017 book on the future of higher education, “Robot Proof: Higher Education in the Age of Artificial Intelligence,” Northeastern University President Joseph Aoun argues that rapid technological advances will require workers to engage in lifelong learning to remain competitive. In his vision, workers would take responsibility for their own development by maintaining a lifelong relationship lifelong universities [xiv]. But there is no reason that workers who already have little to do at work should leave the workplace to gain additional training. With the flexibility afforded by increased automation of work processes, employers could gain a competitive advantage by investing in employee development and continuous education as a core part of the employment experience.
There are a lot of more interesting things that we can do while we’re at work than pretending to work. Already, a measurable portion of the working population is reporting that they aren’t doing all that much particularly productive or useful at work. As technology continues to transform our workplaces, we are likely to continue to find more and more jobs that someone needs to have that don’t require all that much work. If we can’t find a way to get past the fear and embarrassment of admitting that our jobs don’t really require all that much work, we are going to be stuck filling more of our time with bullshit and pretending to work. But if we can talk openly about what our jobs actually demand, we can begin to imagine ways to transform our workplaces to make work a healthier and more empowering experience and allow people to actualize their human potential while participating in wage labor.
[i] John Maynard Keynes, ‘Economic Possibilites for Our Grandchildren’, in Essays in Persuasion (New York: Harcourt, Brace and Company, 1930), pp. 358–73 <https://blackboard.cornell.edu/bbcswebdav/pid-3890351-dt-content-rid-21812309_1/courses/17239_2018FA_COMBINED-COMEET/Keynes1930.pdf> [accessed 12 November 2018].
[ii] Norbert Wiener, The Human Use Of Human Beings: Cybernetics And Society, New edition edition (New York, N.Y: Da Capo Press, 1950).
[iii] Murray Bookchin, Post-Scarcity Anarchism (Berkeley [Calif.]: Ramparts Press, 1971).
[iv] Martin Ford, Rise of the Robots: Technology and the Threat of a Jobless Future (New York, NY: Basic Books, a member of the Perseus Books Group, 2015) <https://cornell-library.skillport.com/skillportfe/main.action?assetid=106952> [accessed 11 December 2018].
[v] Carl Benedikt Frey and Michael A. Osborne, ‘The Future of Employment: How Susceptible Are Jobs to Computerisation?’, Technological Forecasting and Social Change, 114 (2013), 254–80 <https://doi.org/10.1016/j.techfore.2016.08.019>.
[vi] Michael Chui, James Manyika, and Mehdi Miremadi, ‘Where Machines Could Replace Humans — and Where They Can’t (yet) | McKinsey’, 2016 <https://www.mckinsey.com/business-functions/digital-mckinsey/our-insights/where-machines-could-replace-humans-and-where-they-cant-yet> [accessed 27 August 2018].
[vii] Lisa Dodson and Rebekah Zincavage, ‘It’s Like a Family’, in Caring on the Clock: The Complexeties and Contradictions of Paid Care Work (New Brunswick, New Jersey: Rutgers University Press, 2015).
[viii] Edwin A. Locke and Gary P. Latham, ‘Building a Practically Useful Theory of Goal Setting and Task Motivation: A 35-Year Odyssey.’, American Psychologist, 57.9 (2002), 705–17 <https://doi.org/10.1037//0003-066X.57.9.705>.
[ix] Gary Johns, ‘Some Unintended Consequences of Job Design’, Journal of Organizational Behavior, 31.2–3 (2010), 361–69 <https://doi.org/10.1002/job.669>.
[x] Kimberly D. Elsbach and Andrew B. Hargadon, ‘Enhancing Creativity Through “Mindless” Work: A Framework of Workday Design’, Organization Science, 17.4 (2006), 470–83 <https://doi.org/10.1287/orsc.1060.0193>.
[xi] David Graeber, ‘On the Phenomenon of Bullshit Jobs’, STRIKE! Magazine, 2013 <https://strikemag.org> [accessed 20 November 2018].
[xii] David Graeber, Bullshit Jobs, First Simon & Schuster hardcover edition. (New York: Simon & Schuster, 2018).
[xiii] Elizabeth Dreike Almer and Steven E. Kaplan, ‘The Effects of Flexible Work Arrangements on Stressors, Burnout, and Behavioral Job Outcomes in Public Accounting’, Behavioral Research in Accounting, 14.1 (2002), 1–34 <https://doi.org/10.2308/bria.2002.14.1.1>.
[xiv] Joseph Aoun, Robot-Proof: Higher Education in the Age of Artificial Intelligence (Cambridge, Massachusetts: The MIT Press, 2017).