A 15-hour work week. That's what influential economist John Maynard Keynes prophesied in his famous 1930 essay "Economic Possibilities for Our Grandchildren," forecasting that in the next century technology would make us so productive we wouldn't know what to do with all our free time.
This is not the future Keynes imagined.
Many higher income workers put in 50 or more hours per week, according to an NPR/Harvard/Robert Wood Johnson Foundation poll. Meanwhile, lower-income workers are fighting to get enough hours to pay the bills, as shown in a University of Washington report on Seattle's $15 minimum wage publicized this week.
Yet some of today's best minds are making Keynes-like predictions. This month, Apple co-founder Steve Wozniak said robots will one day replace us — but we needn't worry for a few hundred years.
In May, Facebook CEO Mark Zuckerberg told Harvard's 2017 class that increased automation would strip us not only of our jobs but also of our sense of purpose.
Automation. Artificial intelligence. Machine learning. Many experts disagree on what these new technologies will mean for the workforce, the economy and our quality of life. But where they do agree is that technology will change (or completely take over) tasks that humans do now. The most pressing question, many economists and labor historians say, is whether people will have the skills to perform the jobs that are left.
"We are moving into an era of extensive automation and a period in which capitalism is just simply not going to need as many workers," said Jennifer Klein, a Yale University professor who focuses on labor history. "It's not just automating in manufacturing but anything with a service counter: grocery stores, movie theaters, car rentals ... and this is now going to move into food service, too.
"What are we going to do in an era that doesn't need as many people? It's not a social question we've seriously addressed."
Instead of worrying about the mass unemployment a robot Armageddon could bring, we should instead shift our attention to making sure workers — particularly low-wage workers — have the skills they need to compete in an automated era, says James Bessen, an economist, Boston University law lecturer, and author of the book Learning by Doing: The Real Connection Between Innovation, Wages, and Wealth.
"The problem is people are losing jobs and we're not doing a good job of getting them the skills and knowledge they need to work for the new jobs," Bessen said.
Addressing this skills gap will require a paradigm shift both in the way we approach job training and in the way we approach education, he said.
"Technology is very disruptive. It is destroying jobs. And while it is creating others, because we don’t have an easy way to transition people from one occupation to another, we’re going to face increased social disruption," he said.
In this new age, Bessen said, we can't treat learning as finite.
"We need to move to a world where there is lifelong learning," he said. "You have to get rid of this idea that we go to school once when we’re young and that covers us for our career. ... Schools need to teach people how to learn, how to teach themselves if necessary."
Universal basic income
A universal basic income (UBI) has been proposed as one possible solution to the loss of jobs caused by automation. A UBI would give everyone a fixed amount of money, regularly, no matter what. Proponents say not only would it help eradicate poverty, but it would be especially useful for people whose jobs are eliminated by automation, giving them the flexibility to learn new skills required in a new job or industry, without having to worry about how they'd eat or pay rent.
Some also suggest it would breed innovation. In his Harvard speech, Zuckerberg told the audience: "We should have a society that measures progress not just by economic metrics like GDP, but by how many of us have a role we find meaningful. We should explore ideas like universal basic income to give everyone a cushion to try new things."
Several countries are exploring or experimenting with a UBI, including Kenya, Finland, the Netherlands and Canada.
Concerns about automation aren't new
Americans have been worrying about automation wiping out jobs for centuries, and in some occupations, automation has drastically reduced the need for human labor.
In 1900, 41% of American workers were employed in agriculture, but by 2000, automated machinery brought that number down to just 2%, MIT professor David Autor wrote in the Journal of Economic Perspectives in 2015.
The arrival of the automobile ushered out horses, reducing the need for blacksmiths and stable hands.
In the 21st century, computers are increasingly performing tasks humans once did.
But the relationship between automation and employment is complex. When automation replaces human labor, it can also reduce cost and improve quality, which, in turn, increases demand.
Such was the case in textiles. In the early 19th century, 98% of the work of a weaver became automated, but the number of textile workers actually grew.
"At the beginning of the 19th century, it was so expensive that ... a typical person had one set of clothing," Bessen said. "As the price started dropping because of automation, people started buying more and more, so that by the 1920s the average person was consuming 10 times as much cloth per capita per year."
More demand for cloth meant a greater need for textile workers. But that demand, eventually, was satisfied.
When ATMs were introduced in the 1970s, people thought they would be a death knell for bank tellers. The number of tellers per bank did fall, but because ATMs reduced the cost of operating a bank branch, more branches opened, which in turn hired more tellers. U.S. bank teller employment rose by 50,000 between 1980 and 2010. But the tasks of those tellers evolved from simply dispensing cash to selling other things the banks provided, like credit cards and loans. And the skills those tellers had that the ATMs didn't — like problem solving — became more valuable.
When computers take over some human tasks within an occupation, Bessen's research shows those occupations grow faster, not slower.
"AI is coming in and it’s going to make accountants that much better, it’s going to make financial advisers that much better, it’s going to make health care providers that much more effective, so we’re going to be using more of their services at least for the next 10 or 20 years," Bessen said.
These examples, though, are of occupations where automation replaces some part of human labor. What about when automation completely replaces the humans in an entire occupation? So far, that's been pretty rare. In a 2016 paper, Bessen looked at 271 detailed occupations used in the 1950 Census and found that while many occupations no longer exist, in only one case was the demise of an occupation attributed mostly to automation: the elevator operator.
A 2017 report from the McKinsey Global Institute found that less than 5% of occupations can be completely automated.
What's in store
History has taught us a lot about how automation disrupts industries, though economists admit they can't account for the infinite ways technology may unsettle work in the future.
When a new era of automation does usher in major economic and social disruption — which Bessen doesn't predict will happen for at least another 30 to 50 years — it's humans that will ultimately decide the ways in which robots get to change the world.
"It's not a threat as much as an opportunity," he said. "It’s how we take advantage of it as individuals and a society that will determine the outcome."