The Shifts — Great and Small — in Workplace Automation

There have been periodic warnings in the last two centuries that automation and new technology would wipe out large numbers of middle-class jobs.

In the early 19th century, for instance, a group of English textile artisans, known as Luddites, famously protested the automation of textile production by seeking to destroy some of the machines. A century later, concern rose again over “The Automation Jobless,” as they were called in the title of a Time magazine story of February 24, 1961. U.S. President Lyndon B. Johnson even empaneled a Commission on Technology, Automation, and Economic Progress to confront the productivity problem in 1964 — specifically, that productivity was rising so fast it might outstrip demand for labor. The Commission ultimately concluded that automation did not threaten employment, but that didn’t permanently close the case.

Employment displacement concerns are valid and have regained prominence. For instance, in their widely discussed book, The Second Machine Age, MIT scholars Erik Brynjolfsson and Andrew McAfee offer an unsettling picture of the likely effects of automation on employment.

While we can say with certainty that the past two centuries of automation and technological progress have not made human labor obsolete — the employment‐to‐population ratio actually rose during the 20th century as women moved from home to market — past interactions between automation and employment do not necessarily predict the future. There’s no fundamental economic law that guarantees every adult a living solely on the basis of sound mind and good character.

In particular, the emergence of greatly improved computing power, artificial intelligence (AI), and robotics raises the possibility of replacing labor on a scale not previously observed. If this should occur, the primary challenge will be one of income distribution rather than non-employment. How can we ensure that the largest number of people gain from the surge in productivity?

Labor Market Polarization

Automation does, indeed, substitute for labor — as it is typically intended to do. However, automation also complements labor, raises output in ways that lead to higher demand for labor, and interacts with adjustments in labor supply.

The frontier of automation is rapidly advancing, yet the challenges to fully substituting machines for workers in tasks requiring flexibility, judgment, and common sense remain immense. In many cases, machines both substitute for and complement human labor — substituting for workers in routine, codifiable tasks, while amplifying the comparative advantage of workers in problem-solving skills, adaptability, and creativity. Focusing only on what is lost misses a central economic mechanism by which automation affects the demand for labor: It raises the value of the tasks that workers uniquely supply.

The biggest challenge as new technologies emerge is in the types of jobs created and what those jobs pay — not the number of jobs per se. Although automation may not prove the enemy of employment, it may pose a large challenge for income distribution. In the last few decades, a “polarization” of the labor market is emerging in which employment gains accrue disproportionately in jobs at the top and bottom of the distribution. The rapid employment growth in both high- and low-education jobs has substantially reduced the share of employment for “middle-skill” jobs. In 1979, the four middle-skill occupations — sales, office and administrative workers, production workers, and operatives — accounted for 60% of employment. In 2007, this number was 49%, and in 2012, it was 46%. (See “Polarization in the U.S. Labor Market.”)

What specific job changes can we expect going forward? Consider the surprising complementarities between information technology (IT) and employment in banking. When automated teller machines (ATMs) were introduced, their numbers in the U.S. economy quadrupled from approximately 100,000 to 400,000 between 1995 and 2010. You might assume that these machines all but eliminated bank tellers in that interval. But U.S. bank teller employment actually rose modestly from 500,000 to approximately 550,000 over the 30-year period from 1980 to 2010.

IT enabled a broader range of bank personnel to become involved in “relationship banking” as the routine cash-handling tasks of bank tellers receded. Banks recognized that tellers could be both checkout clerks and salespeople, forging relationships with customers and introducing them to additional bank services like credit cards, loans, and investment products. This is an example of a fundamental economic reality that is frequently overlooked: Tasks that cannot be substituted by automation are generally complemented by it.

Job Quantity Versus Quality

Even if automation does not reduce the quantity of jobs, it may greatly affect the quality of available jobs. As the price of computing power has fallen, computers and their robot cousins have increasingly displaced workers in accomplishing explicit, codifiable tasks.

Tasks that have proved most vexing to automate are those demanding flexibility, judgment, and common sense — skills that we understand only tacitly. In particular, two broad sets of tasks have proven stubbornly challenging to computerize.

One category requires problem-solving capabilities, intuition, creativity, and persuasion. These tasks, termed “abstract,” are characteristic of professional, technical, and managerial occupations. They employ workers with high levels of education and analytical capability, and they place a premium on inductive reasoning, communications ability, and expert mastery.

The second category includes tasks requiring adaptability, visual and language recognition, and in-person interactions — which we call “manual” tasks. Manual tasks are characteristic of food preparation and serving jobs, cleaning and janitorial work, grounds cleaning and maintenance, in-person health assistance, and numerous jobs in security and protective services.

Workers in abstract, task-intensive occupations benefit from IT because of strong complementarities between routine and abstract tasks, elastic demand for services, and inelastic labor supply to these occupations over the short and medium term — even though these activities present daunting challenges for automation. IT, then, should raise earnings in occupations that make intensive use of abstract tasks and among workers who intensively supply them.

The Role of Machine Learning

When considering how human labor can complement new technology, I see two distinct paths: environmental control and machine learning. The first regularizes the environment, so that comparatively inflexible machines can function semi-autonomously. In the second approach, engineers develop machines that attempt to infer rules from context, abundant data, and applied statistics.

Some researchers expect that as computing power rises and training databases grow, the machine-learning approach will meet or exceed human capabilities. Others suspect that machine learning will only “get it right” on average, while missing many of the most important and informative exceptions.

My prediction is that employment polarization will not continue indefinitely. While some of the tasks in many middle-skill jobs are susceptible to automation, many will continue to demand a mixture of tasks from across the skill spectrum. For example, medical support occupations — radiology technicians, phlebotomists, nurse technicians, and others — are a significant and rapidly growing category of relatively well-remunerated, middle-skill employment. A significant stratum of middle-skill jobs, combining vocational skills with literacy, numeracy, adaptability, problem solving, and common sense, will persist in coming decades.

This prediction has one obvious catch: the ability of the U.S. education and job-training system (both public and private) to produce the kinds of workers who will thrive in these jobs of the future. In this and other ways, the issue is not that middle-level workers are doomed by automation and technology, but instead that human-capital investment must be at the heart of any long-term strategy for producing skills that are complemented, rather than substituted, by technological change.