Toil and Technology

• Author: James Bessen
• Position: Lecturer in Law at the Boston University School of Law; this article draws on his forthcoming book, Learning by Doing: The Real Connection between Innovation, Wages, and Wealth.

At the Quiet Logistics distribution center north of Boston in the United States, a robot lifts a shelf and transports it through the warehouse to a workstation. There, an employee picks an item from the shelf and places it in a shipping box. Each robot in the distribution center does the work of one and a half humans.

Robots and other technologies are transforming supply chains, tracking items from source to consumer, minimizing shipping time and cost, automating clerical tasks, and more. But are they eliminating the need for human workers, leading to persistent technological unemployment?

Surprisingly, the managers of warehouses and other supply chain facilities report that they have difficulty hiring enough workers, at least enough with the skills needed to use the new technologies. Moreover, they see these skill shortages persisting for the next decade.

New “smart machines” are radically changing the nature of work, but the question is how. Powered by artificial intelligence, new technologies are taking over tasks not only from warehouse workers, but also from white-collar workers and professionals. Automated teller machines have taken over the tasks of bank tellers; accounting software has automated the work of bookkeepers. Now computers can diagnose breast cancer from X-rays and predict survival rates at least as well as the average radiologist.

What, exactly, does this mean for jobs and wages? Sometimes new technologies eliminate jobs overall, but sometimes they create demand for new capabilities and new jobs. In one case, the new machines replace workers overall; in the other, they just displace workers to different jobs that require new skills. In the past, it has sometimes taken decades to build the training institutions and labor markets needed to develop major new technical skills on a large scale.

Policymakers need to know which way technology is headed. If it replaces workers, they will need to cope with ever-growing unemployment and widening economic inequality. But if the primary problem is displacement, they will mainly have to develop a workforce with new specialized skills. The two problems call for very different solutions.

Despite fears of widespread technological unemployment, I argue that the data show technology today largely displacing workers to new jobs, not replacing them entirely. Of the major occupational groups, only manufacturing jobs are being eliminated persistently in developed economies—and these losses are offset by growth in other occupations.

Yet all is not well with the workforce. The average worker has seen stagnant wages, and employers report difficulty hiring workers with needed technical skills. As technology creates new opportunities, it creates new demands as well, and training institutions are slow to adapt. Although some economists deny that there are too few workers with needed skills, a careful look at the evidence below suggests we face a significant challenge building a workforce with the knowledge needed to use new technologies. Until training institutions and labor markets do catch up, the benefits of information technology will be limited and not widely shared.

Automation ≠unemployment

I focus on information technology because this technology has brought dramatic change to a large portion of the workforce. Some people see computers automating work and conclude that technological unemployment is inevitable. A recent study (Frey and Osborne, 2013) looks at how computers can perform different job tasks. It concludes that 47...