AI Job Displacement Hits Clerical Workers Harder Than Predicted

Early predictions about artificial intelligence's impact on the workforce may have missed the mark. While analysts consistently warned that managers, financial analysts, and lawyers faced the highest risk of AI displacement, recent employment data tells a different story.

Economist Gad Levanon analyzed two years of government employment statistics and found that occupations long flagged as vulnerable—including management, engineering, science, and law—have not only continued growing but in some cases accelerated their expansion. Meanwhile, roles not considered especially at risk, such as records clerks, bookkeepers, and customer service staff, are contracting rapidly.

Why it matters

This reversal challenges the methodology behind most AI job-impact forecasts, which ranked occupations by counting how many tasks could theoretically be automated. The real-world pattern suggests that task automation and job elimination follow different paths—a distinction that matters for workforce planning, retraining programs, and policy decisions affecting millions of workers.

Testing the theory against reality

Levanon's methodology went beyond simple headcount comparisons. He measured each occupation's growth against the average growth rate across all industries where that occupation appears. This approach isolated whether clerical roles were declining because they happened to sit in struggling sectors, or because the occupations themselves were shrinking.

The pattern held: clerical work is contracting even within industries where managerial positions are expanding. The decline stems from the occupation itself, not its industrial context.

Why professionals are growing, not shrinking

The gap between prediction and reality reflects how organizations actually deploy AI tools. Rudy DeFelice, executive vice president and global head of AI strategy at Harbor Global, points to a critical distinction: "There is a difference between the tasks a professional performs, and the responsibility that comes with those tasks. AI currently automates portions of these jobs, but it does not automate the responsibility attached to them."

In practice, AI handles routine components of professional work—research, drafting, data analysis—while human judgment remains essential for decisions carrying legal, financial, or strategic consequences. Clerical roles, by contrast, often center on tasks that AI can execute end-to-end without requiring human accountability for outcomes.

Rethinking vulnerability metrics

The findings suggest that conventional AI-risk rankings, built on task-level analysis, may systematically misidentify which workers face displacement. Occupations heavy on routine information processing appear more vulnerable than those requiring accountability, client relationships, or judgment under uncertainty—even when the latter involve many automatable tasks.

The research was first reported by Inc., based on Levanon's published analysis of government employment data through 2026.