New research by MIT economist David Autor finds that since 1980, technology has replaced more U.S. jobs than it has created. It is a shift Autor attributes to an increased rate of automation and a slower rate of augmentation. Augmentation represents scenarios where technology drives the creation of new tasks, ultimately generating new job roles. Conversely, automation refers to instances where machinery takes over human tasks, resulting in job losses.
In this first-ever investigation to look into job losses and job creation driven by technology, Autor and his colleagues analyzed over 35,000 U.S. census job categories and studied every U.S. patent filed since 1920. They developed a new method that used the nature of language in the patent texts to establish links between new technologies and their impact on employment. Previous studies had only managed to focus on job losses from new technologies.
The researchers found that approximately 60% of current U.S. jobs represent new forms of work created since 1940. From 1940 through 1980, jobs such as elevator operators and typesetters were automated, but technology also created a need for more workers in roles like shipping and receiving clerks, and civil and aeronautical engineers. From 1980 through 2018, roles such as cabinetmakers and machinists were overtaken by automation, while new work emerged for industrial engineers and operations and systems researchers and analysts.
Autor’s research shows that automation’s negative effect on employment was more than double in 1980-2018 compared to 1940-1980. Despite these observations, Autor notes, “there’s been no period where we haven’t also created new work”. He also highlights the complex relationship between automation and augmentation, which often occur within the same industries.
The research further indicates that over the past 40 years, technological advances have widened the wage gap in the U.S. with highly educated professionals tending to work in newer fields that have both high and low-income jobs. Autor acknowledges that things like demographic shifts and large-scale consumer demand also drive new work, not just technology.
Looking into the future, Autor believes the potential uses of AI in the workplace could reshape the employment landscape, but predicting its effects at this stage remains difficult. Despite this uncertainty, the researcher is confident that their new methods designed to observe the interplay between technology and employment will provide new insights moving forward.
