The workforce transition and its human toll
AI and automation are taking the world by storm, and the future of work is poised for a major transformation. Is AI a threat to the future of work, or an opportunity for greater prosperity? In just a few short years, many tasks may be handled by machines, upending the labor market and potentially leading to significant job disruptions. But how significant, exactly? The answer may surprise you, and it’s crucial to consider the implications for workers, businesses, and society as a whole. Let’s dive into the exciting and sometimes controversial world of AI and automation, and explore the possibilities for our collective future.
Yuval Harari, a historian at the Hebrew University of Jerusalem, predicts that if the 19th century created the working class, the 21st century will create the useless class. Humans will no longer be exploited, but rather irrelevant, the latter being much worse. With no economic power, masses of humanity will have no political power. Workers will continually have to be re-skilled in order to survive, which will be detrimental to their mental health, and therefore met with resistance. Harari argues the Fourth Industrial Revolution is different then the first three, so this time really is different. Others, such as Elon Musk, believe a Universal Basic Income (UBI) will be necessary for large swathes of society to afford basic necessities.
Many throughout history have predicted technology would eliminate jobs. Alexis De Tocqueville wrote in 1835: “In proportion as the division of labor is more extensively applied, the workman becomes more weak, more narrow-minded, and more dependent. The art advances, the artisan recedes.” The art advanced even faster during the Second Industrial Revolution. Henry Ford’s assembly line broke down jobs into ever-smaller tasks, alienating workers from the end product of their toil. In the 1950s, an autoworker remarked: “I like the old jobs on the mill better than the jobs on the new because you are in control. What you do determines how the pipe goes, and what the machines do. You’re on top of the machines. In the new mill the machines are on top of you.” It was not just factory workers who were impacted, as cyberneticist Norbert Wiener observed: “The degradation of the scientist as an independent worker and thinker to that of a morally irresponsible stooge in a science factory has proceeded….rapidly and devastatingly.”*
John Maynard Keynes predicted technological unemployment “due to our discovery of means of economizing the use of labor outrunning the pace at which we can find new uses for labor.” Workers have been continually deskilled throughout history, so Klaus Schwab wonders: “What should we do to foster more positive outcomes and help those caught in the transition?”
A potential solution is to treat workers displaced by technology the same way the US treats workers displaced by international trade: through Trade Adjustment Assistance (TAA) that pays a stipend for a set period of time and also provides retraining and relocation assistance until the worker is prepared for a new role. In the face of automation in the Third Industrial Revolution, the United Auto Workers (UAW) called for “rehiring, retraining and increased unemployment and welfare benefits.” Today, if a worker can prove that their job loss was a direct result of AI, then they would be entitled to these benefits. The onus comes in proving that your job was in fact lost due to AI; of all requests for compensation sent to the Tariff Commission between 1962 and 1969, zero were approved. Talk of social protection is abundant, but action is rare.
OpenAI CEO Sam Altman proposes an American Equity Plan to cushion US workers caught between the jobs of today and tomorrow. This would give all US citizens a share of GDP. Altman argues: “I believe that owning something like a share in America would align all of us in making the country as successful as possible – the better the country does, the better everyone does…and we should consider eventually replacing some of our current aid programs, which distort incentives and are needlessly complicated and inefficient, with something like this.” The American process of applying for benefits is ripe for disruption, as AI could optimize & streamline a byzantine, stressful and inefficient process.
Whether the solution is UBI, TAA, or AEP, a monetary cushion will avert class-conflict. In the 19th century, Luddites smashed new machines. Daniel Dobrygowski argues they: ‘‘weren’t protesting technology – they were protesting the decisions powerful people were making about technology that threatened to rob them of their careers, violate their rights, and cast them into poverty.” Surely workers would be less fearful of new technology deployment if they knew there was a safety net to catch them.
In the 20th century, Detroit auto worker James Boggs believed those without jobs would hit those who still had jobs over the head in order to get money to eat. Boggs believed that in an equal society technological advancement could lead to a condition whereby everyone could have their milk and honey without toil, focusing their energies on more creative pursuits. Yet he believed greed would prevent this techno utopia from taking root. With the benefit of hindsight, and with unemployment levels at all time lows in the US at the time of writing, Boggs was too far ahead of his time. Where he was right was in warning automation would upskill white workers while pushing black workers to unskilled “scavenger” jobs.
Labor historian Jason Resnikoff argues that factory “automation,” a term coined by Ford Motor Company in 1947, did not eliminate jobs, but rather sped them up. Whereas DeepMind and Inflection AI Co-Founder Mustapha Suleyman argues: “AI is fundamentally a labor-replacing tool,” Resnikoff sees the problem being less with the technology and more with the inherently exploitative nature of capitalism, i.e. class-conflict. Oxford University professor Michael Wooldridge notes: “Computers don’t get tired, don’t turn up to work hungover or late, don’t argue, don’t complain, don’t require unions, and perhaps more important, they don’t need to be paid.” Many may soon discover that their employers are just not that into them.
On the flip side, in an old parable, UAW president Walter Reuther was shown around a Ford plant in Cleveland when a company official pointed to newly installed robots and asked: “How are you going to collect union dues from these guys?” Reuther replied: “How are you going to get them to buy Fords?”* Employees are consumers, after all.
In 1956 the Vatican worried: “Will automation create “ghost towns”? What about the problems of small business in the automated age? Could we again face the problem of poverty in the midst of plenty, as in the 1930s? Can we afford to leave the solution to be worked out by blind economic forces?”* This is not a question of mere antiquarian interest. As I wrote in a past essay, and was backed up by a recent study out of the University of Colorado, job loss is a leading cause of suicide. Governments should not only think about, but also implement, solutions before rather than after the jobs are gone. Otherwise the economic transition may be soaked in blood.
Some say the past is prologue, yet Dr. Harari argues that although history is important, we have never been here before; the past is not coming back, nor would anybody want to go back, and therefore this time really is different.
Jobs may not be eliminated, but like in the past, changed. The World Economic Forum (WEF) has forecasted that 75 million jobs will disappear, while 133 million new roles will be created. In order for future generations to survive, lifelong learning will have to become as natural as walking. Business, government, labor unions, professional associations and academia should work together to continuously re-skill and upskill the workforce. One controversial idea is to tax AI, robots, and corporations, to finance this. Whether workers will be able to keep up is an open question, as the divide between older generations and what Eric Schmidt, Henry Kissinger and Daniel Huttenlocher describe as “AI natives” grows, which is why a multi-stakeholder approach utilizing cutting-edge technologies will be required.
Sugata Mitra argues learning is not something we have to make happen, but rather let happen. The best teachers will instruct students on how to become lifelong learners as opposed to focusing on specific subject matter. The WEF’s Education 4.0 Framework aims to move education “from a system where learning is standardized, to one based on the diverse individual needs of each learner, and flexible enough to enable each learner to progress at their own pace.” Virtual-Reality and emotion recognition may help to accelerate learning. The future belongs to those who embrace change.
Of course none of this will matter if Artificial General Intelligence (AGI) kills us all, as ethicist Eliezer Yudkowsky argues. Meta’s Yann LeCun thinks this view is nonsense. Do not worry, however, for “The AI does not hate you, nor does it love you, but you are made out of atoms which it can use for something else.” Oxford University philosopher Nick Bostrom argues, “A sufficiently intelligent AI won’t stay confined to computers for long. In today’s world you can email DNA strings to laboratories that will produce proteins on demand, allowing an AI initially confined to the internet to build artificial life forms or bootstrap straight to post biological molecular manufacturing.” If Bostrom, LeCun and Yudkowsky cannot agree on what the future will look like, then how could anybody forecast how to prepare the workforce with confidence? Vincent Muller, channeling Socrates, hits the nail on the head:
“There is a serious question whether rigorous work is even possible at this point, given that we are speculating about the risks from something about which we know very little. The current state of AI is not sufficiently specific to limit that space of possibilities enough. To make matters worse, the object of our study may be more intelligent than us, perhaps far more intelligent, which seems to imply (though this needs clarification) that even if we were to know a lot about it, its ways must ultimately remain unfathomable and uncontrollable to us mere humans.”
No teacher can know what skills will be relevant in the coming decades. With more data, the speed of change will accelerate, and therefore the future will become even harder to predict. Programs like Denmark’s National Skills Anticipation Program, which attempts to forecast future trends, will become more important. Others include Singapore SkillsFuture, Denmark Flexicurity, the WEF’s Closing the Skill Gap Accelerator & Global Learning Network, Luxembourg’s Digital Skills Bridge, UNESCO’s Lifelong Learning Cities, HSBC Skills Insight Hub, ManpowerGroup & Welcome.US Jobs Exchange and MIT Solve. More of this is needed, particularly from business, as companies usually know better than the government what their workforce needs will be. James Manyika of the McKinsey Global Institute says: “We’re not convinced that on our current course and speed we’re well set up to manage those transitions in terms of skilling, educating and on-the-job training. We actually worry more about that question of transition than about the ‘Will there be enough work?’ question.”
This sounds nice in theory. Everybody who loses their job due to AI will be retrained, the argument goes, and rapidly placed in jobs that require more in-demand skills with higher wages. But in reality, labor’s experience in the Third Industrial Revolution refutes this claim. During the “Armour Experiment” of 1960, out of 433 packinghouse workers displaced by machines, only 170 were interested in reskilling, 60 were eligible for the program, 58 took the course, and fewer than 20 wound up working jobs using these new skills. In a similar program in Connecticut, 3,500 applicants were screened for retraining, and only 84 completed the course. Who, or what, will be there to catch the majority who fall through the cracks? If history is any guide, it will be hospitals, psychiatric units, prisons and funeral homes.
Sure, a worker could be reskilled to find a new job, but who is going to cover them while they do so? Where will the new job be located? How many workers will actually do it? Will it pay as much? Perhaps if there are no jobs in the future, this is good news. Who would want to work in a dirty factory if they did not have to?
A potential variable now that was not present in the 1960s is the augmentation of human intelligence. In his 2005 book The Singularity is Near, futurist Ray Kurzweil predicts: “Billions of nanobots in the capillaries of the brain will also vastly extend human intelligence.” Google Co-Founder Larry Page accused Tesla Co-Founder Elon Musk of being a “speciest,” to which Musk replied, “Well, yes, I am pro-human, I fucking like humanity, dude.” Sam Altman writes: “We will be the first species ever to design our own descendants.” This raises major ethical issues, as at first only the ultra-wealthy will be able to have these medical procedures for themselves and their offspring, deepening inequality. Those who attempt to forecast the workforce of the future with current human competencies may have the rug pulled from underneath them.
Perhaps “ultraintelligent machines” will be better able to anticipate coming transformations. As far back as 1965 British mathematician I.J. Good wrote:
“Let an ultraintelligent machine be defined as a machine that can far surpass all the intellectual activities of any man however clever. Since the design of machines is one of these intellectual activities, an ultraintelligent machine could design even better machines; there would then unquestionably be an “intelligence explosion,” and the intelligence of man would be left far behind. Thus the first ultraintelligent machine is the last invention that man need ever make, provided that the machine is docile enough to tell us how to keep it under control…It is more probable than not that, within the twentieth century, an ultraintelligent machine will be built and that it will be the last invention that man need make.”
Given humanity’s track record in the first three industrial revolutions, we may be better off letting the machines come up with a social protection plan, using sentiment analysis to build the political will to get it passed through Congress. Nick Bostrom argues: “Once artificial intelligence reaches a level of intelligence that is equal to or greater than that of humans, it will create a self-perpetuating cycle of improvement. AIs would use their intelligence to create even smarter AIs, which would in turn use their enhanced intelligence to create even smarter AIs, leading to a rapid advancement in AI capabilities.” Eric Drexler, when discussing nanotechnology, worried: “Given all that such technology can do, perhaps governments would simply decide that they no longer need citizens!”
For now, this need not necessarily be a free lunch, as benefits could be tied to retraining. COVID-19 did more to change education since the establishment of the University of Bologna a millennium ago. With technological advancement, there is no reason we cannot continue to learn in a personally customized fashion, anytime and anywhere. In 19th century Britain workers moved from the countryside to Lancashire to work in the mills. In the 20th century US, workers traveled from the South to Detroit to toil in the factories. In the Third Industrial Revolution the microprocessor increased output, but shifted jobs away from the urban core. The difference in the Fourth Industrial Revolution is that much retraining can be done remotely. AI is more likely to impact skilled, rather than unskilled, labor. This may even affect politicians, which is reason to hope governments will be more responsive
The Global South will experience job loss as 3D printing technology transfers production from say Bangkok to Boston. Reskilling and upskilling should be an international objective, otherwise disparities will continue to widen. Public-private partnerships toward this aim and Industry 4.0 knowledge centers should become more prominent. Otherwise the Fourth Industrial Revolution will benefit the rich at the expense of the poor, the young at the expense of the old, the latter of whom will be particularly reluctant to reskill, and the large corporations to the detriment of small and medium-size enterprises, where “SMEs employ 60 to 70% of workers in OECD countries and exceed 80% of job creation in some emerging economies.”
It will also be important to create innovative ways of recording, transferring and verifying skills, perhaps baking microcredentials into interoperable Digital ID’s using zero-knowledge proofs (ZKP’s) on the blockchain, or even the iris of our eyeballs. Some sort of universal and transferable language for skills can increase labor mobility.
Some readers may wonder why even work? If robots and AI can do much of what we are currently paid for, why not share the benefits with all? My fear is that the capitalist ethos is baked so deeply into our DNA that even if we only need a fraction of the current workforce to keep the economy moving, society will create new tasks just to keep people busy. The owners of the means of production would rather get something out of us as opposed to splitting pie. This is nothing new, as the late David Graeber taught us, but suffice to say the world will keep moving whether or not we have social media managers or HR consultants. Structures are laid bare when they weaken, as we learned during COVID-19 that many jobs are “non-essential.”
Folks would then have to find meaning outside of work, perhaps devoting more time to sport or study. Masses of humanity would become free of work, only to realize that “freedom sucks,” as David Brooks argues in The Second Mountain. Obligation is what gives life meaning. Futurist Martin Ford suggests we pay people to read; what better way to create an informed citizenry?
Predicting future unemployment rates is a complex task, influenced by a range of factors including population growth, workforce demographics, technological advancements, and economic trends, not to mention war & peace. How many people will be seeking work? How many hours a week will they work? What age will workers retire (a big question given the expanding human lifespan)? What age will people enter the workforce? What about immigration? It is impossible to forecast mass unemployment without also looking at birthrates, which are in decline across aging Western societies.
Society may need to slow the rate of expansion, shorten hours or reduce the number of years people work to adjust. Roboticist Rodney Brooks argues: “We have overpopulated the planet, so we have to go this way to survive. I’m very worried about the standard of living dropping because there’s not enough labor as I get older.” To accurately forecast unemployment rates and prepare for the future of work, it’s essential to consider a wide range of factors and to continually adapt to new developments and trends.
Today, just like in the past, workers are resistant to new technologies. Most could be forgiven for their hesitance, for if they lose their job, at least in the US, there is hardly any safety net to catch them. Spending years harnessing a craft and then starting from scratch takes a toll on the psyche, which is why emotional healthcare should be part of any “just transition.” Better yet, how can we make it to such assistance would not be needed?
How does extraterrestrial labor fit into all of this? I am joking; we have enough queries to answer. Though most humans cannot control the future, we can ask the right questions. Perhaps learning how to do so, with machines taking care of much else, will be the most important skill we can acquire. While a photo can be beautiful, a painting by a human artist has a unique style and emotion that a camera simply cannot capture.
—
* I visited the Walter P. Reuther Archive of Labor and Urban Affairs in Detroit, Michgan, USA for these quotes. If you are interested in knowing the exact collection, box and folder for any particular citation, please reach out.
Author’s note: it would be impossible for me to survey the opinions of every authority on a question as big as the future of work, so this is by no means meant to be a comprehensive review. If there are contrarian opinions I missed, but you believe are valuable for future essays, feel free to drop me a line.