After robots take over our jobs, what is next? Guaranteed incomes? A public service job creation program? More conferences on “the future of work”?
Today, not a week goes by without a new study being released or conference convened on robots replacing jobs or a newspaper account of a new use of robots taking the place of workers (this week Domino’s Pizza announced it was using robots to deliver pizzas in Europe). In turn, the subject has come center stage in meetings of education and workforce professionals. At the March meeting of the California Workforce Development Board (CWDB), for example, several board members asked whether California state government was lagging in its policy responses to and its recognition of job loss through automation.
Should the state of California be doing more? Should the national government? What should be the responses of workforce and education professionals, like those on the CWDB, and other policymakers, public and private?
Below are three key lessons we’ve learned in the past half century about automation, job loss and the options now available to workforce practitioners and policymakers:
1– Since the 1960s, nearly all warnings about automation and higher unemployment have proved incorrect (though this time could be different). In the mid-1960s, our state government in California was concerned about automation eliminating jobs — very concerned — and established a series of policy groups and task forces to consider whether the state was entering a period of ongoing unemployment over 10 percent or more. This followed a series of actions by the national government fearing automation, including the creation of an Office of Automation and Manpower in the Department of Labor in 1961 and the appointment of the National Commission on Technology, Automation and Economic Progress in 1964.
Similar automation concerns since the mid-1960s have arisen from time to time. When I entered the job training world in the early 1980s, the unemployment rates in California were rising quickly, above 8 percent and higher, and state policymakers were convinced that de-industrialization and the decommissioning of our state’s heavy manufacturing was leading to permanently high unemployment. In the early 1990s, various task forces related to automation were established in response to the demise of the state’s aerospace industry, and later in the early 2000s, the internet was the villain in feared job losses among retail clerks, bank tellers and other service workers.
In each period, of course, automation did eliminate jobs, lots of jobs. But more jobs were created in occupations and sectors not even envisioned a few years earlier. Just as most economists had predicted since the 19th century, as automation improved productivity, buying power increased and new jobs were created. Economists continue to take this view, perhaps none more so than David Autor of M.I.T, who discusses the ongoing process of technology destroying and creating jobs in this TED talk on employment from September 2016.
Yet, just because automation and technology in the past two centuries have resulted in net job gains, does not mean this will continue in the future. For a long time Harvard economist Lawrence Summers dismissed employment concerns about automation, until the past few years, when he has begun to openly worry that this phase of automation will be different. In March, two of the most prominent researchers on automation, Daron Acemoglu of M.I.T and Pascual Restrepo of Boston University issued a paper, expressing caution about their earlier projections of continued net job gains.
Whether automation will continue to yield net job gains (and how these jobs are distributed geographically) remains to be determined. In the next few years, our understanding will benefit from more field reports on automation (like this thoughtful report on automation and employment in the warehousing field in California by Natalie Kitroeff of the Los Angeles Times) as well as attempts to track employment impacts across the economy.
2– The workforce system in the United States has succeeded well over the years in adapting to technology and automation, though going forward it will be challenged to increase its response time. The first job training program I worked on in 1982, at a community job training program in San Francisco, trained workers in the repair of typewriters and copiers. Business machine shops on nearby Howard Street were hiring repairpersons, as were Xerox and other major copier companies. Within a few years, all of these shops were out of business, and the copier companies had moved on to new technologies.
The training program quickly adapted, and moved on to train computer repairpersons and technicians in the growing occupations of microwave technician and telecommunications technician. Other training programs of the time for bank tellers, secretaries and automotive mechanics also adapted and either moved to new expanding occupations or (more often) rewrote training curriculum to reflect new technologies needed by bank tellers, secretaries or automotive mechanics.
The workforce system in the United States, now centered on the hundreds of local Workforce Development Boards throughout the country, has been the subject of criticism for as long as I can remember—for being out of touch and not training for jobs in demand. In fact this has never been the case, and is not the case today. The local Boards, in the great majority of cases, stay close to the local labor markets, know the local employers, are pragmatic and focused on job placements.
Bob Lanter has been in the workforce system for more than two decades, as a director of a local Board, and currently as the executive director of the California Workforce Association, the association of the 46 local Boards in the state. “For my members, on-going contact with employers is a main priority, probably the main priority. The Boards are constantly being pushed to be pro-active, to keep up with and even get ahead of the skills that technology is requiring.”
As Lanter notes, going forward, workforce practitioners will be challenged to pivot and adapt likely at even faster paces. “When my members get together, a frequent topic is the increased speed of technological change, and how we’re going to respond in real-time to revising training designs.”
3– Policy entrepreneurs are rushing forward to advance guaranteed income schemes and other “end of work” schemes; experience suggests we be cautious, very cautious. As the research and discussion on automation have grown, so also have the “end of work” schemes put forward by policy entrepreneurs.
The guaranteed income idea — the idea that as the total number of jobs decreases, government should guarantee a certain income to all citizens — has gained greatest traction. Its proponents include libertarian Silicon Valley venture capitalists, as well as welfare state expansionists on the left. The guaranteed income is advocated with a variety of rationales — simplifying the complex and costly government benefit systems, responding to the rise of the contingent work of Uber and TaskRabbit — but compensation for job elimination is the main rationale.
There have been a few experiments in the United States with forms of income maintenance. In the late 1960s and early 1970s, experiments with forms of income maintenance were undertaken in Seattle and Denver, but were abandoned by the mid-1970s, without any significant follow-up. Canada has been the site of guaranteed income pilots, including Mincome during the early 1970s in Manitoba, another experiment that ended after a few years. Finland started a pilot of income maintenance for 2,000 citizens earlier this year.
There are a lot of practical issues with guaranteed income, starting with costs. There are even greater issues regarding the quality of life without structured employment. Though the concept of women and men freed from work has long been a goal of social reformers, research on laid-off workers in recent years has found that men in particular do not spend their time in fulfilling enterprises of craft, caring for family members or poetry. Instead, they are in a depressed condition of watching television, sleeping and surfing the internet and more socially isolated than ever.
Beyond guaranteed income are other ideas for addressing job loss and automation — new forms of tax credits, large-scale public service jobs programs even new communities of artisans in Columbus, Ohio. Brookings Center of Technology director Darrell West has put forward a wide range of other ideas to address automation, including incentives for volunteerism, lifelong learning and expanding the earned income tax credit. All of these, though, are forms of government income maintenance, with their own costs and distances from structured employment.
Ultimately, how we think about responses to automation will be rooted in our views on the role of employment for individuals and for the broader society. Some among us will celebrate the movement away from the structure and repetition of most jobs, and favor forms of income maintenance. Others (among whom I count myself) will see in structured employment the hedge against individual and social disorders, and seek to continue its roles. These are the big political debates that have gone on for centuries and especially since technology brought the promise of widespread affluence and leisure in the 19th century. We will continue to debate them in the years ahead, as our affluence and opportunities for leisure continue to mount.
In the meantime, we want to be weary of the inflated industry of “future of work” theorists (whose main products are speculative conferences, papers and being paid to meet among themselves) even as we look to more serious researchers and journalists to help us better understand the current and next phases of automation.