SHRM report: AI exposure rises, but high displacement risk falls to 5.1%

Artificial intelligence and automation affected more tasks across U.S. workplaces over the past year, but the share of wage-and-salary employment facing high displacement risk declined to 5.1%, according to 2026 research released June 25 by the Society for Human Resource Management.

SHRM’s estimate translates to about 7.9 million jobs, using May 2025 employment estimates from the Bureau of Labor Statistics. That figure is a modeled assessment of potential exposure and displacement risk. It is not a count of jobs already lost to AI, and it does not mean that 7.9 million workers have been laid off.

What the report measures

The research is based on SHRM’s 2026 Automation/AI Survey and occupational data covering the U.S. wage-and-salary workforce. The survey was conducted in spring 2026. It examines how work tasks are being automated, how occupations are exposed to automation and which occupations may face a high risk of displacement.

SHRM reported that average task automation increased over the prior year. At the same time, its estimate of the share of employment at high automation-displacement risk fell from 6% in the prior estimate to 5.1% in the new research.

Those findings describe two related but different developments. More tasks may be handled or assisted by automated systems, while the estimated share of jobs facing the highest level of potential displacement can still decline. Occupational exposure also does not establish that any particular worker will lose a job.

White-collar exposure is not universal

The report identifies white-collar occupations as generally more exposed to automation. It also notes exceptions, including occupations in construction and extraction. That distinction matters because exposure is not distributed evenly across the workforce, and a broad occupational pattern cannot predict the outcome for every job or employer.

SHRM’s practical conclusion is that AI is increasingly transforming human labor rather than eliminating it overall. In that framing, the immediate workplace effect is more likely to involve changes to tasks and the need for reskilling than the disappearance of most jobs.

That conclusion does not eliminate the possibility of job losses. SHRM says specific occupations may experience displacement, even as the overall research points toward greater transformation of work. The 5.1% estimate therefore identifies a substantial group of workers and jobs that may face elevated risk without establishing that the risk has become an actual employment loss.

How to read the 7.9 million figure

The job total is a conversion of the 5.1% estimate using May 2025 BLS occupational-employment data. It should be understood as an estimate of the number of jobs in the high-risk category, not as a real-time count of AI-caused layoffs or a measure of total U.S. employment.

The evidence in the approved report does not establish how these findings compare with the BLS employment release scheduled for July 2026. That means the SHRM research provides a risk assessment and a snapshot of changing task automation, but it does not by itself show whether overall employment has risen or fallen because of AI.

For workers and employers, the report’s central message is mixed: automation exposure is rising, but the estimated share of employment at high displacement risk is lower than in the prior estimate. The next broader employment data will provide a separate measurement of jobs and employment; the SHRM report itself does not announce a specific policy, employer deadline or follow-up action.