This is the second of two parts MIT News Based on new research by Ford Economics Professor David Autor, the special study examines the number of new jobs created in the United States since 1940. Part 1 is available here.
Ever since the Luddites destroyed the looms, it’s been clear that new technology kills jobs. But technological innovation also creates new jobs: think of computer programmers, or people installing solar panels on their roofs.
Overall, is technology displacing more jobs than it creates? What is the net balance between these two things? So far, this has not been measured. But a new research project led by MIT economist David Autor has found the answer, at least for American history since 1940.
The study used new methods to examine how many jobs have been lost to machine automation, and how many have been created through “augmentation” (where technology creates new tasks). Research has found that online, especially since 1980, technology has replaced more U.S. jobs than it created.
“Over the past 40 years from 1980 to the present, automation appears to have occurred faster and augmentation slower than in the previous 40 years,” said Autor, a co-author of a newly published paper. Author. Detail the results.
However, this discovery is just one of the advances in this research. The researchers also developed a novel method to study the question, based on an analysis of tens of thousands of U.S. Census job categories and a comprehensive look at U.S. patent texts from the past century. This allows them to quantify the impact of technology on unemployment and job creation for the first time.
Previously, scholars had largely been able to quantify job losses, not job gains, caused by new technologies.
“I feel like a paleontologist looking for the dinosaur bones that we thought must be there but couldn’t find them until now,” Otto said. “I think this study is a breakthrough on something we suspected was true, but before this study we had no direct evidence.”
The paper “New Frontiers: The Origins and Content of New Works, 1940-2018” appears in quarterly journal of economics. Co-authors include Autor, Ford Professor of Economics; Caroline Chin, a doctoral student in economics at MIT; Anna Salomons, professor at Utrecht University School of Economics; Bryan Seegmiller SM ’20, PhD ’22, associate at Northwestern University’s Kellogg School professor.
Automation and enhancement
Overall, about 60 percent of jobs in the United States are new types of jobs created since 1940, the study found. A century ago, computer programmers might have been working on farms.
To determine this, Otto and his colleagues combed through some 35,000 job categories listed in U.S. Census Bureau reports, tracking how they emerged over time. They also used natural language processing tools to analyze the text of every U.S. patent filed since 1920. The study examined how words are “embedded” in census and patent documents to mine relevant text passages. This allows them to identify links between new technologies and their impact on employment.
“You can think of automation as a machine that takes work input and does the work for a worker,” Otto explains. “We think of augmentation as a technology that increases the variety of things people can do, the quality of things people can do, or their productivity.”
For example, from about 1940 to about 1980, jobs such as elevator operators and typesetters tended to be automated. But at the same time, more workers are filling positions such as shipping and receiving clerks, buyers and department heads, and civil and aerospace engineers, and technology creates the need for more workers.
From 1980 to 2018, the ranks of people such as cabinet makers and machinists shrank due to automation, while the ranks of people such as industrial engineers and operations and systems researchers and analysts grew.
Ultimately, the study shows that the negative impact of automation on employment was more than twice as large in the 1980-2018 period as in the 1940-1980 period. Compared to 1940 to 1980, the impact of increases on employment changed more modestly and positively between 1980 and 2018.
“There’s no law that says these things have to be balanced one-to-one, although we create new pieces all the time,” commented O’Toole.
What will artificial intelligence do?
However, the study also revealed many nuances in the process, as automation and augmentation often occur within the same industry. Technology has not only reduced the number of farmers, it has also created air traffic controllers. For example, within the same large manufacturing company, there may be fewer machinists but more systems analysts.
Relatedly, technology trends over the past 40 years have exacerbated the wage gap in the United States, with highly educated professionals more likely to work in new fields, which themselves are divided into high- and low-paying jobs.
“The new jobs are bifurcated,” O’Toole said. “As old jobs are removed in the middle, new jobs are growing on both sides.”
Research also shows that technology is not the only factor driving new jobs. Demographic changes are also behind the growth of many sectors of the service industry. Interestingly, new research also shows that large-scale consumer demand also drives technological innovation. Inventions are not only provided by smart people who think outside the box, but also in response to clear social needs.
Eighty years of data also show that the future path of innovation and its impact on employment is difficult to predict. Consider the possible uses of artificial intelligence in the workplace.
“AI is really different,” Otto said. “It may replace some highly skilled expertise, but it may complement decision-making tasks. I think we are in an era where we have this new tool but don’t know what use it is for. New technology has its pros and cons and needs It took a while to figure it out. GPS was invented for military purposes and took decades to make its way into smartphones.”
He added: “We hope our research methods will allow us to learn more about future developments.”
As Autor acknowledges, there is room for further refinement of the research team’s approach. For now, he believes the study opens up new areas of research.
“The missing link is documenting and quantifying the extent to which technology enhances people’s jobs,” Autor said. “All the previous measures just showed automation and its impact on displaced workers. We were surprised that we were able to identify, classify and quantify the enhancements. So to me, that in itself is very basic.”
This research was supported in part by the Carnegie Corporation; Google; the Gak Institute; the work of the MIT Futures Task Force; Schmidt Futures; the Smith-Richardson Foundation; and the Washington Center for Equitable Growth.