Robots aren’t killing jobs fast enough—and we should be worried

The US economy needs to learn new tricks.
The US economy needs to learn new tricks.
Image: Reuters/Kim Kyung Hoon
We may earn a commission from links on this page.

Never has there been more discussion of the disruption of work and, by some measures, never has there been less disruption.

It could be the calm before the AI-aided storm. It could also be a sign of a stagnating American economy.

“Take a deep breath and calm down,” write the authors of a recent study on how technology will impact jobs. Futurists and Silicon Valley-types worried about robots and artificial intelligence taking over for humans are overly alarmed, they say, adding that it is misleading when people like the World Economic Forum’s Klaus Schwab say that “the speed of current breakthroughs has no historical precedent.”

In fact, data show that the US labor market is the calmest it has been in more than 160 years. The problem is there is not enough disruption. If anything, we need more jobs destroyed.

That argument, made by Robert Atkinson and John Wu of the Information Technology and Innovation Foundation, a think tank promoting policies that spur innovation, is against the grain.

They make a compelling case

Their belief that we are in an age of stagnation, not disruption, is based on a decade-by-decade analysis of how quickly occupations have been appearing and disappearing since 1850. The methodology involves summing up the share of percentage increase or decrease in the number of jobs in every occupation category, as defined by the US Bureau of Labor Statistics, to find the total percentage of jobs created and destroyed. By this method, the 1950s was an extremely disruptive period, with a job-churn rate at a 20th-century high of 37%.

While the types of jobs people did changed drastically from 1950 to 1970, very little changed from 2000 to 2015. The  rate in the 2000s was an all-time low of 14%, and thus far, the 2010s are even lower, at 6%.

Job disruption in late 19th and 20th centuries was mostly attributable to the introduction of agricultural technologies that reduced the need for farm workers. The number of household workers also declined as low-cost appliances like washing machines and vacuum cleaners replaced domestic help. Jobs in manufacturing and the service sector replaced these jobs.

Today, the composition of jobs is relatively stagnant. Atkinson and Wu say this is because true innovation is at an ebb. New digital technologies are impressive, the authors say, but not nearly as revolutionary as the changes that arose from widely available electricity and steel in the 1890s, and the electromechanical technologies that emerged in the 1950s and 1960s that greatly increased manufacturing efficiency. The lack of new technology has hurt labor productivity, the researchers argue, and it is one of the reasons wages have risen so little since 1980. We need policies to promote disruptive new technologies like those used by Uber and Amazon.

It may still be worth preparing for the worst

Andrew Stern, former president of the Service Employees International Union, says Atkinson and Wu’s argument is thought provoking, yet it is no reason to stop being worried about the future. “There are an enormous number of academics and people creating the future who think this could be a big problem,” Stern told Quartz.

We should take those people seriously, he believes, and at the very least “plan for the scenario” where the job apocalypse arises. Stern also points out that the study does not take into account the falling number of working-age people who have jobs, the falling share of US GDP that goes to workers, and the rising number of jobs that are insecure—most of the job growth since 2005 has been in temporary positions or independent contracting.

Ethan Pollack, an economist at the Aspen Institute’s Future of Work Initiative, shares Stern’s belief that it is worth preparing for the worst. Pollack says there is evidence that certain technologies, like AI, could take off and have a massive effects on jobs. The US is not good at helping workers get back on their feet, so this kind of job destruction would have devastating effects.

“It’s been decades since the waves of automation and outsourcing hit the manufacturing communities in the Rust Belt,” Pollack told Quartz. “And we’re still dealing with the fallout from that.”