The Industries Most Exposed to AI Are Adding Jobs, Not Cutting Them

The intuitive story about AI and employment goes like this: industries that deploy AI at scale are the ones where jobs will disappear first. It is a reasonable assumption. It is also, according to a study published last week, largely wrong, at least so far.

Christos Makridis, an associate research professor at Arizona State University, and economist Andrew Johnston analyzed employment data covering nearly all US employers from 2017 to 2024. They supplemented that with Gallup Workforce Panel data from more than 30,000 employees collected through 2026. Their finding: industries with the highest AI exposure saw 3.9% more job growth and 4.8% higher wages than comparable sectors, alongside 10% productivity gains.

That is not what the job-destruction narrative predicts. But it is consistent with how previous waves of automation played out, where sectors that adopted new tools fastest tended to grow output, which then required more workers, not fewer.

What the numbers show

The study finds that each percentage-point increase in frequent AI users within a state or industry correlates with 0.1 to 0.2% higher real output and 0.2 to 0.4% higher employment. Between mid-2024 and late 2025, the share of frequent AI users doubled, from 12% to 26%. Extrapolating from those coefficients, that shift alone accounts for 1.4 to 2.8% higher real output across the economy.

The sectors seeing the strongest effect are ones where AI functions as a force multiplier rather than a replacement: marketing, financial analysis, and writing-intensive roles. In those fields, experienced workers appear to benefit most, with wages rising in occupations that place a high value on tacit knowledge and judgment. Entry-level workers face more displacement pressure, but the study finds that over time those roles tend to be redefined upward, toward tasks that require human oversight and exception handling rather than routine execution.

Boston Consulting Group reached a similar conclusion in its own 2026 analysis, finding that AI is more likely to reshape jobs than eliminate them outright. "AI has acted as a productivity-enhancing tool that boosts employment and wages rather than a simple substitute for labor," Makridis wrote in summarizing the findings for Singularity Hub.

What this does not resolve

This is not a case against worrying about AI and work. Makridis and Johnston find that the benefits concentrate in states with efficient labor markets. Workers in states with less labor market flexibility see smaller gains, and there is no guarantee the sectors adding jobs are accessible to the people being displaced elsewhere.

A separate analysis estimates that about 12% of current jobs fall into a category where AI directly substitutes for the core tasks humans do, not augments them. For those workers, the aggregate picture offers cold comfort. Goldman Sachs data published here last week documented 16,000 net US jobs lost per month to AI, with Gen Z entry-level workers bearing most of the cost. The ASU study and the Goldman data are not contradictory. They are measuring different things. The macro picture allows for displacement in some segments and growth in others. The question of who bears the costs, and whether they have access to the sectors that are hiring, is where the analysis gets harder.