August 6, 2026
Upjohn Institute President Michael Horrigan recently returned to the U.S. Bureau of Labor Statistics (BLS), where he spent more than 30 years overseeing the nation’s employment, unemployment, and inflation measurement programs, to present research, jointly with Erik Vasilauskas on artificial intelligence (AI) and the future of work to approximately 260 agency staff members.
The presentation drew on AI Exposure and the Future of Work: Tasks, Skill Demand, and Education, a paper co-authored by Horrigan and Upjohn Research Analyst Erik Vasilauskas. The study examines which workplace activities are most susceptible to AI and digital automation and how employment outcomes vary among occupations with similar levels of exposure.
The presentation was particularly relevant to BLS staff as the research relies extensively on BLS occupational employment data and projections, as well as the U.S. Department of Labor's O*NET database.
Horrigan and Vasilauskas evaluated 2,087 Detailed Work Activities from O*NET. Activities involving documentation, data analysis, computers, scheduling, and information processing ranked among the most exposed to AI. Physical work, caregiving, sensory judgment, and in-person interaction generally ranked much lower.
Exposure is concentrated in knowledge-intensive fields. Nearly all employment in legal occupations and business and financial operations occupations falls within the highest-exposure quartile, along with about three-quarters of employment in computer and mathematical occupations.
However, exposure should not be interpreted as a forecast of job loss. BLS projects employment growth across all four exposure quartiles from 2024 to 2034. Within the most highly exposed occupations, projected outcomes differ sharply by the education typically required for entry:
- Occupations requiring a bachelor's degree or higher are projected to grow 7.2 percent and account for nearly 40 percent of total projected job growth.
- Occupations requiring some college or an associate's degree are projected to decline 4.2 percent.
- Occupations requiring a high school diploma or less are projected to decline 3.6 percent.
The findings suggest that AI may complement workers in some knowledge-intensive professions while assuming a larger share of routine clerical and transactional work. They also underscore the importance of responsive career services and employer-driven training partnerships that help workers adapt as skill needs change.