Workforce Automation: Better Data Needed to Assess and Plan for Effects of Advanced Technologies on Jobs (GAO Report GAO-19-257)

The Policy

What it does

Evaluates the impact advanced technologies have had on the US workforce and compiles select firms’ feedback on the risks and benefits of adopting advanced technologies.


This report was produced by the US Government Accountability Office (GAO) in response to Congress’s request for an investigation into how advanced technologies are impacting the US workforce. GAO surveyed available data and concluded that better data are “needed to assess and plan for the effects of advanced technologies on jobs.” The main finding of this report is that the government is not adequately tracking how advanced technologies are affecting the workforce, but that this information could be helpful to the government, employers, and job seekers. To address this, GAO recommended that the Department of Labor (DOL) use existing or new data to “systematically track the workforce effects of advanced technologies.” The DOL agreed with GAO’s recommendation and will work to fill gaps in the existing data and continue to research this topic.

GAO found that none of the available federal data sets sufficiently explain how or if changes in the workforce are connected to advanced technologies. Therefore, GAO used available federal survey data (American Community Survey (2010-2016), the Current Population Survey’s Displaced Worker Supplement (2016), and the Occupational Employment Statistics (2017)) to look for trends in occupations that had been identified as being susceptible automation; likewise, it interviewed sixteen firms to find examples of how adopting advanced technologies is affecting the US workforce. GAO found no meaningful differences in job loss rates in occupations or industries susceptible to automation. However, these industries did increase their concentration of tech workers during GAO’s study period (2010-2016), “possibly an indicator of industries preparing to adopt advanced technologies.”

GAO found some trends from their interviews with firms adopting advanced technologies. Interviewed firms reported reducing cost, making more consistent products, increasing worker safety, and decreasing tedious tasks as motivators for adopting advanced technologies. For firms that decided to use advanced technologies, some employment trends were observed. There were increases in tech jobs, but the employment shifts were not universally in that direction. At the selected firms, “workers changed roles and tasks as a result of advanced technology adoption, such as focusing more on interactive, cognitive, higher-skilled, and monitoring tasks, and in other cases focusing more on lower-skilled tasks.” Several firms were decreasing the number of employees, but many were doing so by not replacing employees that quit or retired rather than with lay-offs. Additionally, firms often re-trained current employees whose positions were replaced rather than hiring from outside the firm. Worker adaptability was highlighted as essential with the changing demands of firms that adopt advanced technologies.

Additionally, GAO pointed out that jobs susceptible to automation are not equally distributed among education levels, ethnicity, or geographic areas, so some social groups will likely be affected more than others resulting from these changes. Better data on how advanced technologies are affecting the US workforce could help the government better support workers and provide relevant training opportunities.


There are serious concerns about how jobs in the US will be changed or displaced due to automation. A 2017 study by McKinsey Global Institute found that “activities most susceptible to automation involve physical activities in highly structured and predictable environments, as well as the collection and processing of data. In the United States, these activities make up 51 percent of activities in the economy accounting for almost $2.7 trillion in wages.” Additionally, this study estimates that “half of today’s work activities could be automated by 2055.”

GAO has been producing reports on workforce automation for years, including these two reports from the 1980s, in response to requests from Congress. Bobby Scott (Chairman of the House Committee on Education, and Labor), Suzan DelBene (Member of the House Committee on Ways and Means), and Richard Neal (Chairman of the House Committee on Ways and Means) requested that GAO “study the effects of advanced technologies on the US workforce,” which resulted in the current report.

US government agencies responsible for monitoring the US workforce include the Department of Labor (DOL) and the Department of Commerce (DOC). The DOL’s Bureau of Labor Statistics works to identify projected trends in the US workforce, but it does not yet collect data on all the types of advanced technologies used in different occupations, nor does it thoroughly track changes through time.

GAO concluded: “Without comprehensive data that link technological changes to shifts in the workforce, DOL lacks a valuable tool for ensuring that programs it funds to support workers are aligned with local labor market realities, and employers and job seekers need to rely on other sources of information to decide what training to offer or seek.” The DOC’s Census Bureau has begun asking about businesses’ uses of advanced technologies in their Annual Business Survey and is expanding these efforts with questions about why companies adopted advanced technologies and how it has impacted their business and workers. However, this survey will not address the magnitude of workforce changes and, at the time GAO produced this report, the survey was not yet available.

Another study published in 2017 by the National Academies of Sciences, Engineering, and Medicine (NASEM) investigates the current state of technology and work, projections for the years to come, and how these trends fit with the country’s values. NASEM found that, with momentous changes on the horizon, “the outcomes for the workforce and society at large depend on our choices. Technology can be a powerful tool. What do we want for our future society? How do we decide this?” This report, similarly to GAO’s report, concludes that prioritizing data collection and analysis to monitor workforce and technology trends is key.

As of February 2020, there have been several recent related legislative efforts (including HR 2542: PLACE Act, S 1558: Artificial Intelligence Initiative Act, HR 2432: Future DATA Act, and HR 3388: SELF DRIVE Act) to address or investigate workforce automation. Additionally, GAO is currently expanding its efforts in science and technology and plans to double the size of its workforce in this area. Part of the work of the new Science, Technology Assessment, and Analytics (STAA) team will be related to automation. GAO’s strategic plan identified “a number of technologies and scientific advances that will potentially transform society, among them genome editing, artificial intelligence and automation, quantum information science, brain-computer interfaces and augmented reality, and cryptocurrencies and blockchain.”

The Science

Science Synopsis

For the purposes of this report, GAO used “advanced technologies” as an intentionally broad term “to describe technological drivers of workforce changes.” These technologies include those identified in a 2017 study by the National Academies of Sciences, Engineering, and Medicine, namely artificial intelligence, machine learning, robotics, autonomous transport, 3D printing, advanced manufacturing, advanced materials, computing power, and internet and cloud technology.

Workforce automation has been happening in various forms for centuries, from the stocking frame knitting machine of 1589 to Ford’s 1913 Model T assembly line. Researchers have found, through these examples and others, that advanced technologies of the past often replaced skilled labor. Beginning in the twentieth century, computers changed this trend. As computers became more powerful and less expensive, telephone operators were no longer needed, reservations could be made online, ATMs could be used to withdraw cash, and self-checkout machines moved into grocery stores. According to Dr. Michael Osborne and Dr. Carl Frey of Oxford University, “the result has been an increasingly polarised labour market, with growing employment in high-income cognitive jobs and low-income manual occupations, accompanied by a hollowing-out of middle-income routine jobs.” They suggest that there is going to be a shift towards more creative and social occupations.

Scientific Assumptions

Automation will impact many sectors of the economy: The premise of this report is that “robots, artificial intelligence, and other advanced technologies are changing the workplace.” Data suggest that employment shifts are, at least in part, related to automation and the advanced technologies that are being developed.

Frey and Osborne’s assessment of jobs’ automation risk was accurate: GAO analysis of workforce data was based on a peer-reviewed article by Carl Frey and Michael Osborne, “The Future of Employment: How Susceptible Are Jobs to Computerisation?” In this study, 702 jobs were considered and assigned a risk of computerization based on attributes such as their need for social intelligence or creativity. The authors found that approximately 47% of total US employment is at high risk of automation; however, the academic community has not settled the debate on this topic yet and various estimates have been published.

The firms the GAO interviewed are representative of the US workforce: By interviewing select firms, GAO was able to compile some anecdotal evidence of the workforce changes occurring. GAO claims that “the selected firms varied in size, industry sector, types of technologies used, and geographic location” and though they concede that the results are not generalizable, the number and type of firms used to assess trends in the US workforce affects the conclusions. GAO did not provide further information about why these specific firms were selected.

The Debate

Scientific Controversies / Uncertainties

Another prominent study, “A Future That Works: Automation, Employment and Productivity” by the McKinsey Global Institute, analyzed 2,000 work activities as they related to 800 occupations. Though they used different methods, overall the two studies had similar findings about the types of jobs that are most susceptible to automation and the pervasiveness of automation.

There is uncertainty about the size of the workforce disruption from new advanced technologies relative to past workforce shifts. Despite the large changes projected, some think the changes are relatively small and the fears about advanced technologies are unwarranted. While the transition period may be difficult for some groups, there is cause for optimism about the aggregate impact of advanced technologies on the workforce. Over time, technology tends to create new jobs and improve working conditions.

It remains unclear how much of an effect automation will have compared to other workforce trends, and if it is getting a disproportionate amount of attention. For instance, companies have also been transitioning jobs from full-time staff to contractors, which can decrease the salary and benefits associated with the position.

Potential Impacts

GAO suggests that more complete data could help the government, in its efforts to monitor and address workforce changes, and job seekers, as they attempt to adapt to the changing workforce landscape. In response to the report, the DOL conveyed that “The Department is committed to understanding the potential impact of advanced technologies on the workforce; developing new, relevant, and comprehensive measures on the nature of employment; and aligning workforce development programs to continue to meet the needs of employers and workers.”

GAO raised concerns about automation disproportionately affecting workers with lower education levels: “these workers will be in greater need of programmatic or policy supports, and federal workforce programs will need to be aligned with in-demand skills for the changing economy.” Representative Suzan DelBene (D-WA-1), one of the members of Congress who requested this report, has introduced legislation to establish a “Lifelong Learning and Training Account” to help low- to moderate-income workers gain new skills. A similar bill was also introduced multiple times in the Senate by Senator Mark Warner (D-VA). Neither version of this legislation has been passed.