A Problem With No Automatic Solution

Shopping isn’t what it used to be. Go into almost any major retail outlet, and the experience is the same: shelf after shelf and rack after rack of merchandise and not a salesperson in sight. Soon there may not be a cashier in sight either.

The relentless drive to minimize costs by cutting workers is being dramatically accelerated through advances in artificial intelligence (AI) across a range of industries. At risk are such common jobs as cashiers, taxi and truck drivers, bank tellers, agricultural workers, and more.

Even truck driver, a quintessential American occupation, is at risk of extinction; and there is no environmental protection agency that will protect it by labeling it an endangered species. Overall, according to one major study, 47 percent of workers in America had jobs at high risk of potential automation.

Nor is the “autounemployment” threat limited to the blue collar worker, or the less educated, semi-skilled person who works with his hands. The common denominator of “autounemployment” is the routineness of the tasks involved in the job. Thus, white-collar office workers are also at risk. Take radiologists, for example: The latest computerized system made expert human radiologists look like novices, racking up a 50 percent better rate at identifying malignant tumors.

Economists and sociologists are warning that this time around, the new technology will not generate enough new work to make up for the jobs they eliminate. That’s because the automation is so ubiquitous, across so many industries, that the displaced workers will have nowhere to go where their skills and experience will be marketable.

So the question is asked again, and now more urgently than ever: How is society to deal with a coming social crisis triggered by the scourge of mass redundancy? When there is a proliferation of machines that can do almost anything better than humans, what will humans do?

There are two answers. The pessimists say that training and expanded unemployment coverage will help us get through; the optimists say that as in the past, the AI will create work as fast as it displaces people.

The optimists cite data which indicate that current fears are overblown, and in fact technology will provide its own solution. For example: Bank tellers looked to be on the way out, thanks to automated teller machines (ATMs). In the U.S., their average number fell from 20 per bank branch in 1988 to 13 in 2004. Yet, those same ATMs reduced costs so much that it became profitable to open more branches, such that during the same period urban banks expanded by 43 percent.

A similar phenomenon occurred in the occupation of legal research. Legal clerks and paralegals cannot compete with the capacity of computers to survey large quantities of material in preparation for cases. But those same search engines made the process so much more efficient that judges are more inclined to encourage the document-hunting phase known as “discovery,” and the number of legal clerks in America increased by 1.1 percent a year between 2000 and 2013.

Such findings are encouraging, but the majority of experts belong to the doomsaying chorus. It would be folly to ignore their warnings and just have faith that technology will make everything come out all right.

The warnings have been coming at least as far back as 1964, when a panel of Nobel prizewinners, known as the Ad Hoc Committee on the Triple Revolution, sent President Lyndon Johnson a memo alerting him to the danger of social upheaval triggered by “the combination of the computer and the automated self-regulating machine.”

A White House report just out became the latest addition to the shelf of literature on the subject. It recommends that the federal government expand both access to education in technical fields and the scope of unemployment benefits.

Whether the incoming Trump administration and its congressional majorities will act on that could depend on the future of the economy. As of now, the Republicans are aiming at a trimming back of entitlement programs to reduce government spending. The goal of fiscal responsibility is laudatory, but if self-driving cars start to put taxi drivers out of work, they will need not only training but also financial assistance to help them over what will be a painful period of adjustment.

Congress will likely be tempted to hold lengthy hearings on the issue, and eventually issue their findings in a thick volume that few will read. But in reality, we already have the studies on this. What’s needed now is action.