MIT News

Ever because the Luddites destroyed power looms, it has been clear that latest technologies can destroy jobs. But technological innovations also create latest jobs: consider a pc programmer or someone installing solar panels on a roof.

Overall, is technology replacing more jobs than it creates? What is the web balance between these two things? This has not been measured thus far. But a brand new research project led by MIT economist David Autor has developed a solution, not less than for U.S. history since 1940.

Using latest methods, the study examines what number of jobs have been lost through machine automation and what number of have been created through “augmentation,” by which technology creates latest tasks. The study finds that technology on the Internet, particularly since 1980, has replaced more jobs than it has created within the United States.

“In the last 4 many years from 1980 to the current, there appears to be a faster rate of automation and a slower rate of expansion than within the 4 many years before,” says Autor, co-author of a newly published paper detailing the outcomes.

However, this result’s just one in every of the study’s advances. The researchers have also developed a wholly latest method for studying the issue, based on an evaluation of tens of hundreds of U.S. Census occupational categories combined with a comprehensive take a look at the text of U.S. patents from the last century. This allowed them to, for the primary time, quantify the impact of technology on each job loss and job creation.

To date, researchers have largely only been in a position to quantify job losses brought on by latest technologies, not job gains.

“I feel like a paleontologist on the lookout for dinosaur bones that we thought existed but have not been in a position to find,” says Autor. “I feel this research brings latest insights into things that we suspected were true but for which we had no direct evidence before this study.”

The paper, “New Frontiers: The Origins and Contents of New Works, 1940-2018“, appears in . The co-authors are Autor, the Ford Professor of Economics; Caroline Chin, a graduate student in economics at MIT; Anna Salomons, Professor on the School of Economics at Utrecht University; and Bryan Seegmiller SM ’20, PhD ’22, assistant professor on the Kellogg School at Northwestern University.

Automation versus expansion

The study finds that overall, about 60 percent of U.S. jobs represent latest varieties of work created since 1940. A century ago, this computer programmer could have worked on a farm.

To discover, Autor and his colleagues combed through about 35,000 job categories listed in U.S. Census Bureau reports and tracked how they emerged over time. They also used natural language processing tools to investigate the text of all U.S. patents filed since 1920. The research examined how words were “embedded” in census and patent documents to disclose related passages of text. This allowed them to discover connections between latest technologies and their impact on employment.

“You can consider automation as a machine that takes input from a job and does it for the employee,” explains Autor. “We consider augmentation as a technology that increases the range of things people can do, the standard of things people can do, or their productivity.”

For example, from about 1940 to 1980, jobs comparable to elevator operators and typesetters tended to turn out to be automated. At the identical time, nevertheless, more employees filled positions comparable to shipping and receiving clerks, buyers and department managers, and civil and aeronautical engineers, where technology created a necessity for more employees.

From 1980 to 2018, the variety of carpenters and machinists, amongst other things, was thinned out by automation, while the number of business engineers and operations and systems researchers and analysts, for instance, increased.

Ultimately, the research suggests that the negative impact of automation on employment within the period 1980-2018 was greater than twice as large as within the period 1940-1980. Compared to 1940-1980, there was a more modest and positive change within the impact of job creation on employment between 1980 and 2018.

“There is not any law that these items need to be balanced one-to-one, although there has never been a time when we have not also created latest work,” Autor notes.

What will AI do?

However, research also uncovers many nuances on this process, as automation and expansion often occur inside the same industries. It’s not only that technology is decimating the ranks of farmers while creating air traffic controllers. Within the identical large manufacturing company, there could also be fewer machinists but more systems analysts.

In this context, technological trends over the past 40 years have exacerbated the wage gap within the United States, as well-educated professionals usually tend to work in latest fields which might be themselves divided between high-paying and low-paying jobs.

“The latest work is split into two parts,” says Autor. “While old work was deleted in the center, latest work emerged on each side.”

As research also shows, technology shouldn’t be the one driver of recent work. Demographic change can be the reason for the expansion of various sectors of the service industry. Interestingly, the brand new research also suggests that prime consumer demand can be driving technological innovation. Inventions are delivered not only by smart individuals who think outside the box, but in response to clear societal needs.

The 80 years of information also suggests that future paths for innovation and the impact on employment are difficult to predict. Consider the possible uses of AI within the workplace.

“AI is absolutely different,” says Autor. “It may replace some high-level expertise, but it may possibly also complement decision-making tasks. I feel we’re in a time where we’ve got this latest tool and we do not know what it’s good for. New technologies have strengths and weaknesses and it takes some time to figure them out. GPS was invented for military purposes and took many years to seek out its way into smartphones.”

He adds: “We hope that our research approach will allow us to say more about this in the longer term.”

As Autor acknowledges, there may be room for further refinement of the research team’s methods. Currently, he believes the research opens up latest study opportunities.

“The missing link was documenting and quantifying how technology was improving people’s jobs,” says Autor. “All previous measures simply highlighted automation and its impact on workforce displacement. We were amazed that we were in a position to discover, classify and quantify augmentations. So that in itself is pretty fundamental to me.”

The research was supported partially by the Carnegie Corporation; Google; Gak Institute; the MIT Work of the Future Task Force; Schmidt Futures; the Smith Richardson Foundation; and the Washington Center for Equitable Growth.

This article was originally published at