Automation has been a feature of technological change throughout human history, particularly since the start of the Industrial Revolution. However, rather than progressing at a regular, predictable pace, automation often occurs unpredictably during short bursts of rapid technological change.
Given that we currently find ourselves in such a period, there is renewed interest in understanding the types of jobs that are susceptible to displacement through automation, along with the likely number of workers affected. Many estimates have been produced over the years, but consensus remains elusive.
In response to this uncertainly, SHRM has embarked on new research in 2025 that aims to estimate the extent to which automation and related technologies (e.g., generative AI) will displace human labor in the U.S. workforce. As a first step, we have developed a novel approach to measuring automation displacement risk among U.S. workers
This research uses data from the Occupational Information Network (O*NET), a government database designed to track the characteristics of hundreds of occupations. Our approach relies on the O*NET “degree of automation” data, which reports the extent to which individual occupations are currently automated (i.e., not at all, slightly, moderately, highly, or completely automated).
We assume that an occupation’s current level of automation is predictive of future automation displacement risk. Furthermore, we allow automation displacement risk to vary within an individual occupation. For example, if 20% of O*NET respondents in a given occupation report that their job is highly or completely automated, we project that 20% of employment in that occupation faces a high or very high risk of automation displacement.
Many Will Be Affected, But Fewer Replaced
Occupational Employment and Wage Statistics (OEWS) data for May 2023 from the U.S. Bureau of Labor Statistics (BLS) reports a total nonfarm national employment level of about 151.9 million people. We estimate that a great majority of this employment (87.4%) is currently not at all, slightly, or moderately automated (see Figure 1). This suggests two notable things:
- Most jobs currently require a sufficient level of human input to make their full displacement by automation tools unlikely in the immediate future.
- Nearly half of all jobs currently involve a slight or moderate level of automation, which suggests that advances in automation technology will transform a huge fraction of U.S. employment, even if relatively few jobs are displaced.
Although we estimate that most jobs are unlikely to be displaced by automation in the near term, this risk does exist. Specifically, we estimate 19.2 million jobs (12.6%) are currently highly or completely automated, putting them at high or very high risk of near-term automation displacement. This figure comprises about 16 million jobs (10.5%) that we estimate face high risk and 3.2 million jobs (2.1%) that face very high risk.
Which occupations will feel the most impact? Exposure to high or very high automation displacement risk is not evenly distributed across occupational groups (see Figure 2).
On the low end of the spectrum, we estimate that less than 8% of employment is exposed to high or very high automation displacement risk in six occupation groups, including less than 5% in one case (life, physical, and social science occupations). Individual occupations in these groups often place strong emphasis on characteristics such as critical/creative thinking, creative expression, interpersonal interaction, and/or nonroutinized tasks.
On the high end, we find that at least 15% of employment faces high or very high automation displacement risk in four occupational groups, including nearly 20% of employment in business and financial operations. Often, occupations in these high-risk groups involve a large amount of routinized labor done by machines and software.
What Lies Ahead?
Before interpreting these findings, it’s critical to note that our estimate of 19.2 million jobs facing high or very high automation displacement risk should not be construed as a projection of job losses stemming from automation. Instead, this research provides a sense of scale and highlights which jobs are most vulnerable. SHRM will continue to refine our estimates to assess automation’s impact on job loss, creation, and transformation.
That said, the findings of this initial research provide some important insights for HR executives regarding the likely evolution of work, especially for organizations in which workers’ jobs are already highly automated. Plus, it’s important to note that complete displacement of workers by automation isn’t purely a technology issue. In many cases, the automate-or-not question depends on legal and regulatory issues, or is affected by factors such as customer preferences for human labor.
As HR executives navigate this era of rapid automation, the key challenges are not just anticipating displacement and replacement but actively shaping the future of work and focusing on transformation of roles. HR leaders must focus on workforce agility by investing in continuous learning, reskilling, and redesigning roles to complement automation rather than compete with it.
Justin Ladner is the senior labor economist at SHRM.