Why fears of an AI job-killing apocalypse are overstated

June 12, 2019

When a new generation of industrial robots started arriving in Rust Belt towns in the early 1990s, factory workers panicked. According to the National Bureau of Economic Research, each reprogrammable factory robot put in service between 1990 and 2007 eliminated six human jobs.

That’s a scary statistic and it’s no wonder that workers today are worried about robots and artificial intelligence (AI) taking their jobs. In the past, workers displaced by new technology could always find new work. Farmers could move to factories; factory workers could take office jobs. But there’s a legitimate concern that today’s technological advances will result in not just short-term pain as workers move to new industries, but leave some workers without real long-term options.

While no one can foresee exactly how this will play out, a mountain of evidence suggests that just like during past technological leaps, the fears — though realistic — can be managed. The lesson of history — throughout the centuries and in modern times — is that scary stories about job-stealing machines reliably lead our human minds astray, focusing us on the short-term pain caused by new technologies while minimizing the even greater long-term benefits.

Economists David Autor and Anna Salomons see one source of this cognitive error. The negative effects of new technology tend to be easy to spot — and emotionally resonant. When a machine replaces a human, a real person loses their job, often at significant cost to themselves, their family and their community. At the same time, the economists note, “the indirect effects of technological progress … are likely more challenging to observe and quantify, and hence may receive short shrift in economic analysis and in the wider public debate.”

To quantify these harder-to-measure benefits, Autor and Salomons studied 19 industrialized countries, comparing employment changes in specific industries with the rest of the country’s economy. In country after country, the same pattern emerged: As technology caused some people’s jobs to disappear in one industry, less-noticeable indirect effects led to an overall net increase in employment in the country.

“Aggregate employment grew dramatically in all countries from 1970 to 2007, even as relative employment fell in the industries experiencing the fastest productivity growth,” they said. “We find that technological progress is broadly employment-augmenting in the aggregate.”

Yet, it seems that the natural human tendency has always been to assume the opposite — that this time and this technology must be different. That was true way back in 1598, when Queen Elizabeth I recoiled at the demonstration of a new machine that would speed the manufacture of silk stockings: “I have too much love to my poor people, who obtain their bread by the employment of knitting, to give my money to forward an invention which will tend to their ruin, by depriving them of employment, and thus make them beggars.” (Two centuries later, the same fears drove textile workers known as Luddites to burn factory equipment in protest, the most famous example of panic over job-killing technology.)

Another part of the exaggerated fear of technology-induced unemployment comes from overestimating the speed with which new technologies will remake entire industries, says Paul Oyer, a professor of economics and senior fellow at the Stanford Institute for Economic Policy Research. People overlook the delay between the time the new machine or process is invented and when it actually hits the factory floor. “The workplace isn’t evolving as fast as we thought it would,” Oyer said during a panel discussion on AI job effects early this year.

But will we spend that extra time wisely?

The next transition requires what all the others did: improvements in how we educate kids and adults looking to transition into a new kind of employment. This is the one place where we could screw things up. While losing a job to outsourcing can feel horrible, it’ll be a much bigger problem if there are employment opportunities that people can’t fill because of the lack of proper training. It’s up to us to protect the most vulnerable from direct impact. Just how to go about that is the critical question—and the subject of my next column.

By Douglas Merrill

Source: Forbes