SHRM's high-AI-use estimate is not a headcount forecast

Keep four workforce measures in separate columns

The direct answer is to maintain separate fields for AI-tool use, task automation, nontechnical barriers, and an actual staffing decision. AI use asks whether technology participates in work; automation asks whether a task proceeds without human input; a barrier asks whether law, customer preference, trust, physical presence, accountability, or another condition still requires people; displacement concerns elimination of filled jobs. A role can show high AI use and low automation when employees use tools to augment judgment. It can also show substantial automation while legal or client conditions preserve human work. Combining the four into one exposure score erases the distinctions SHRM's report was designed to surface.

Translate national occupation evidence into local tasks cautiously

Use the SHRM estimates to prioritize interviews and task observation, not to assign a percentage to a named employee or local role. For each selected job family, record the current tasks, frequency and criticality, AI tools used, degree of human intervention, input and output quality, exceptions, decision rights, required relationships, legal and policy constraints, customer expectations, accessibility needs, workload, and skills. Reconcile the organization's job architecture with the survey occupation definitions and document mismatches. SHRM uses a broad AI definition and notes that respondents may report high AI use because AI features are intertwined with computer tools, which can overstate AI's central importance in some contexts.

Require operating evidence before changing work or roles

A workforce proposal should identify the specific task change, responsible process owner, affected employees, expected capacity or quality effect, failure and exception work, training, redeployment path, accommodation, employee and manager input, customer consequence, control owner, review period, and reversal condition. Test the proposed workflow under normal volume, peak volume, poor inputs, unavailable tools, policy exceptions, and handoffs that still require tacit knowledge. Compare measured time, quality, rework, incidents, customer impact, worker experience, and newly created monitoring or coordination work. A national estimate cannot show whether a buyer's licensed product, data, management practice, labor agreement, or local process can sustain the assumed change.

Make headcount a governed decision, not a survey inference

Before any hiring freeze, role elimination, span change, or redeployment, require a dated decision record with local task evidence, alternatives, financial assumptions, legal and employee-relations review, workforce and customer effects, skills supply, transition support, responsible executives, and monitoring. Challenge whether apparent capacity is durable, whether demand will grow, whether quality or control work moves elsewhere, and whether nontechnical barriers were measured rather than assumed away. Report SHRM's 21%, 20%, 60.4%, and 5.1% estimates with their different definitions and population. The CHRO can use the research to make local workforce planning more precise; it cannot make the employment decision for the organization.

Turn this source into a reviewable decision

For AI for CHROs, use this briefing as a dated decision record rather than a substitute for the source. Preserve Automation, AI, and Job Displacement Risk in U.S. Employment | SHRM, the exact URL, the August 25, 2026 review date, the supported facts above, the editorial interpretation, the limitations, and any buyer-specific evidence. Link that record to the decisions most directly affected: Workforce and capacity planning; People analytics and employee listening; Skills intelligence and internal mobility; Workplace AI governance and change. State whether the source changes the scope, evidence requirement, control, sequence, or only the language used to describe the decision.

Before action, name the accountable owner, affected population and workflow, exact offering or configuration, source data and rights, human decision point, exception and appeal path, complete cost, expected benefit, failure and stop conditions, retained evidence, and next review date. Keep official facts, provider statements, buyer observations, representative tests, measured outcomes, editorial inferences, and unknowns visibly separate. Reopen the record when the source, offer, model, integration, data, policy, population, responsible person, or measured result changes.

Limitations and unknowns

SHRM is the research publisher. Its 2026 report describes an April survey of 14,245 U.S. workers, occupation-level estimates linked to May 2025 BLS wage-and-salary employment, broad definitions of AI and automation, a 21% high-AI-use estimate, a 20% high-task-automation estimate, a 60.4% nontechnical-barrier estimate, and a 5.1% high-displacement-risk estimate. The detailed methodological appendix is described as forthcoming, and the research does not establish a particular employer's tasks, tools, barriers, workforce decisions, causality, or future outcomes. Current methodology and data, employer job and task records, employee and manager evidence, representative workflow tests, alternatives and transition analysis, and qualified people, operations, finance, accessibility, privacy, security, employee-relations, labor, and legal review control.