THE ROBOTS are coming! And they are going to take all our jobs! That’s not the message of Friday’s employment update from the Labor Department, which showed that the market for labor is still encouragingly brisk eight years into a moderate but steady economic recovery. Rather, it is a summary of an emerging conventional wisdom about the future of work, which holds that advances in technology, especially artificial intelligence, will soon enable “brilliant” machines to replace every sort of worker from car-wash attendant to appellate lawyer.
To be sure, such ideas have a long history, going all the way back to England’s Luddites, the early-19th-century weavers who smashed mechanical looms in defense of their jobs. But the revolutionary technology of our times, from driverless cars to voice-command computing, makes the potential impact seem especially threatening — even if alarms about a “robocalypse” may be overblown. Certainly some of the world’s leading economists are thinking hard about the future of work, if any; and the latest research received a hearing at no less august a venue than the recently completed European Central Bank conference in Sintra, Portugal. To summarize a paper presented there by economists David Autor of MIT and Anna Salomons of Utrecht University: There’s some reassuring news and some challenging news.
The reassuring part is that productivity growth of the kind associated with rapid, labor-saving technological advancement still results in overall higher employment. Yes, employment declines in one sector — manufacturing, say — but the resultant positive “spillover effects” spawn new investment opportunities in new industries, which create new jobs. Mr. Autor and Ms. Salomons found that this dynamic prevailed across 19 developed countries, the United States included, from 1970 through 2007. It possibly grew less robust after the turn of the 21st century, the economists note, but the robocalypse is still not a short-run likelihood.
More challenging, however, is their finding that, in recent decades, employment growth has not been evenly distributed but rather “polarized,” in the sense that the United States and other developed economies are producing large numbers of jobs for highly skilled workers and less-skilled workers — but relatively few in between. This, in turn, has fed rising income inequality. A clear policy implication, then, is that societies redouble their efforts to provide education and training, both to young people preparing to enter the labor market and to adults who are already in it but must constantly update their capabilities to compete. At the same time, governments should enact mildly redistributive policies, such as the earned-income tax credit wage subsidy, that augment earnings for those who are willing to work but whose skill levels might not otherwise gain them enough wages to sustain a family.
Today’s worry about mass technological unemployment is nothing new and probably won’t come to pass in its scariest imaginable form. Yet that’s no reason for complacency. The life prospects of flesh-and-blood working people can’t just be left on autopilot.