This is not the first time society has fretted over the impact of ever-smarter machines on jobs and work—and not the first time we have overreacted. In the Depression-beset 1930s, labor Jeremiahs warned that robots would decimate American factory jobs. Three decades later, mid-1960s prognosticators offered a hopeful silver lining to an otherwise apocalyptic assessment of automation’s dark cloud: the displacement of work and workers would usher in a new “leisure society.”
Paul Saffo teaches forecasting at Stanford University and chairs the Future Studies and Forecasting track at Singularity University.
Reality stubbornly ignored 1930s and 1960s expectations. The robots of extravagant imagination never arrived. There was ample job turbulence but as Keynes forecast in 1930, machines created more jobs than they destroyed. Boosted by a World War, unemployment dropped from a high of 25 percent in 1933 to under two percent in 1944. And the hoped-for 1960s leisure society never arrived because the diffusion of information technologies created unprecedented demand for Drucker’s “knowledge workers,” and fueled the arrival of the service economy.
Now the specter of job-killing robots is back. A coincidence of factors from a jittery post-crash global economy to dot.com disruption and the relentless advance of Moore’s Law has a new generation of prognosticators pitching dark warnings and the prospect of radical change. Jobs will evaporate and work will cease to be what gives us income—and meaning. Material goods will become abundant to the point of costlessness, and nations will pay citizens a guaranteed minimum income.
Even in the face of today’s considerable uncertainty, it is a safe bet that the most extravagant claims are no more likely now than they were in 1965 or 1933. We are headed toward neither apocalypse nor nirvana. Uncertainty will abound, but change will be far less radical than predicted, and events will unfold slowly enough for society to adapt, albeit painfully at moments.
We will muddle through in the undramatic middle. But amid the muddling, short-term responses will do much to shape long-term outcomes. With this in mind, what follows is a reality check of some of the issues being debated at the intersection of technology and work.
We need a new vocabulary: We have shrunk room-sized computers to the size of dust motes and sent robotic spacecraft past Pluto, but we still use UNIVAC-era vocabulary when talking about new work challenges. Consider that hapless 1940s neologism, automation. Coined by a retired Ford executive to describe the narrow phenomenon of automatic machines deployed on automotive assembly lines (“auto” + “mation”—get it?), automation has become a meaningless catch-all for a much larger revolution. We desperately need a new and richer vocabulary, or at least need to choose among existing terms with greater care. Cybermation anyone?
Let’s not abandon Keynes just yet: In 1930, Keynes observed that technological unemployment was a self-solving problem. On balance new technologies create more jobs than they destroy. Today’s job-shedding turbulence looks no different from what scared the bejesus out of observers in the 1930s and ’60s. For example, in 1965 the federal government reported that automation was wiping out 35,000 jobs per week, yet, just a few years later, it was clear that new jobs more than offset the losses. Of course, now as then, the new jobs will arrive more slowly than the old jobs are destroyed, and require ever-higher skill levels. We would be wise to worry less about extreme scenarios and focus on managing the transition.
Follow the new scarcities to the new jobs: Every new abundance creates a new scarcity that in turn leads to new economic activity. The proliferation of computers made information abundant, creating the demand for Drucker’s knowledge workers. And the material abundance made possible by machine-enabled productivity gains in turn contributed to the rise of an economy hungry for service workers. This moment is no different; immediate job losses are highly visible, while entirely new job categories run beneath the radar. Jobs will be ever less secure, but work isn’t disappearing.
Guaranteed Minimum Income—don’t hold your breath: Guaranteed Minimum Income—paying individuals whether or not they work—is a fascinating concept. But outside its limited use as a small-scale experimental alternative to welfare or development funds, history suggests it is unlikely in the extreme. The technocrats espoused a similar idea in the 1930s, and in the mid-’60s, a group self-identified as the Ad Hoc Committee on the Triple Revolution (cybernation, atomic weaponry, and human rights) proposed that the government “provide every individual and every family with an adequate income as a matter of right.” What we got instead was Reaganomics and, eventually, the One Percent. Particularly in the United States, with its myth of the work ethic and its aversion to “moral hazard,” Guaranteed Minimum Income will be as hard a sell as Leninism.
Watch for the jobs that are never created: The drama of jobs lost is irresistible, be it elevator operators in the 1950s, telephone operators in the ’60s, longshoremen in the ’70s, or truck drivers facing robot competition today. But if the wild card of a jobless future arrives, it will be because of jobs never created to begin with. Consider Facebook: when it went public in 2012, it reported annual gross revenues of $3.7 billion ($1 billion net), accounted for 12 percent of Internet traffic (more than Google), was adding 1.5 million users per day—and had barely 2,400 employees. The same pattern can be seen across cyberspace, from Airbnb to Twitter and Uber. The global population is growing—merely keeping the jobs that already exist isn’t going to put everyone to work.
Velocity matters: Without a doubt, the world of work is in for profound change over the next several decades. Work won’t disappear, but its nature will likely change beyond all recognition, just as it has over the last 70 years. The open question is whether this will be a wrenching shift or something less turbulent. The problem with predictions is that would-be seers tend to condense the time dimension, arguing that vast change will happen overnight. I do not doubt the scale of change ahead, but I am certain that it will happen no more rapidly than the changes in work over the last century. We will marvel at the shifts, but they will happen at a rate that is manageable—if we are wise.
In short, plenty will change in the world of work over the next few decades, but apocalypse is unlikely. Instead, the pattern will be a familiar extension of what already has unfolded in the last century. Jobs will be less secure, an ever greater portion of the workforce will be unwilling independent contractors, and the notion of pursuing a single career will seem as quaint as receiving a gold watch upon retirement. This provides little comfort to workers facing under-employment or worse, but it means that managing the transitions ahead is well within the capabilities of institutions, governments, and societies as a whole. Provided that we find the collective will to do so.
For the Future of Work, a special project from the Center for Advanced Study in the Behavioral Sciences at Stanford University, business and labor leaders, social scientists, technology visionaries, activists, and journalists weigh in on the most consequential changes in the workplace, and what anxieties and possibilities they might produce.