AI Job Risk Hits Knowledge Workers First, Studies Show

Artificial intelligence is reshaping the employment landscape in ways that diverge sharply from past automation waves, according to new research that identifies which workers face the greatest risk over the next several years.

A study from Tufts University's Digital Planet initiative ranked 784 U.S. occupations across 20 industry sectors, measuring vulnerability based on AI's current trajectory. The findings reveal a labor market paradox: the roles most enhanced by AI assistance are often the same ones most threatened by displacement.

Why it matters

Unlike previous automation that targeted manufacturing or routine clerical work, generative AI directly competes with high-skill cognitive labor—the foundation of knowledge economies. Organizations relying on writers, analysts, and junior technical staff will need to rethink workforce planning as AI becomes capable of core professional tasks once considered automation-proof.

Cognitive roles show highest exposure

The Tufts analysis found writers and authors face 57% risk exposure, computer programmers 55%, and web designers 55% over the next two to five years. Software developers, management analysts, and market researchers represent the largest potential income losses due to their combination of high salaries and workforce size.

Bhaskar Chakravorti, dean of global business at Tufts' Fletcher School and a lead researcher, explained the counterintuitive dynamic to CNBC, which first reported these findings. High-tech workers perform exactly the kind of analytical and creative tasks AI systems are rapidly mastering, making those roles vulnerable even as the technology boosts individual productivity for workers who remain.

Junior workers face steeper declines

Research from Stanford's Digital Economy Lab using ADP payroll data covering millions of workers identified a striking pattern: early-career employees ages 22 to 25 in AI-exposed occupations saw employment drop 16% relative to peers, while older workers in identical roles held steady.

Erik Brynjolfsson, director of the Stanford lab, attributes this gap to the nature of what AI replaces versus what it complements. The technology substitutes for formal knowledge that recent graduates typically offer, while augmenting the tacit judgment that experienced professionals develop over time.

The employment effects concentrate in roles where AI automates work or directly substitutes for junior-level contributions. In positions where AI serves as an assistant rather than replacement, entry-level hiring has stabilized or even grown.

Task disruption, not wholesale elimination

Both research teams emphasize that entire occupations are unlikely to vanish. Brynjolfsson stressed that even highly exposed roles contain numerous tasks AI cannot perform. Understanding the shift requires analyzing specific tasks rather than broad job categories.

Health care illustrates this nuance. Despite high salaries, physicians including cardiologists and psychiatrists show lower exposure in the Tufts rankings. Chakravorti noted that AI will augment rather than displace most healthcare professionals, potentially freeing capacity to serve more patients in the same timeframe.

Adoption remains early stage

Brynjolfsson cautioned that current labor market signals represent only the leading edge of change, as most workers have not yet fully integrated generative AI into their workflows. He emphasized that outcomes will depend on deliberate choices by companies, policymakers, and workers rather than technological determinism alone.

The research underscores that while industrialization ultimately created far more jobs than it destroyed, the transition spanned decades during which wage growth lagged productivity gains. This AI-driven wave is moving considerably faster.

These findings were originally reported by CNBC, drawing on studies from Tufts University's Digital Planet initiative and Stanford's Digital Economy Lab.