The Pay Report
Occupations more exposed to generative AI have seen larger increases in advertised salaries than less-exposed ones, according to a September 17 Indeed Hiring Lab analysis by economist Jack Kennedy. Advertised pay in the most AI-exposed occupations has climbed by about 46% since 2021, versus 25% in the least-exposed. These are changes in salaries listed in job ads, not raises every worker received, and not proof that AI caused the gap.
Examples of high-exposure occupations in the analysis include software development, IT support, data analytics, marketing and finance. Low-exposure examples include nursing, personal care, food service, cleaning and manufacturing.
The premium shrinks as the comparison tightens
Indeed tests whether the gap survives when similar jobs are compared. With occupation and month controls, the estimated premium for high exposure was about 5.7%. Controlling instead for job title and month brought it to about 4.7%. When occupation-by-seniority and month controls were used, the estimate fell to 2.4% and was not statistically significant at the conventional 5% level.
That sequence matters. High-exposure occupations also tend to differ in seniority, credentials and responsibilities, all of which affect offered pay. Once those differences are accounted for, most of the apparent AI premium disappears.
Seniority is where the gap lives
The analysis finds a larger divergence in advertised pay growth between high- and low-exposure jobs for senior roles than for entry-level ones. That pattern could reflect the mix of jobs being advertised rather than the technology used in them. Posted salaries also leave out jobs that do not disclose pay, and they do not capture what workers negotiate after an offer.
For context, Indeed’s September snapshot reported that posted wages overall rose 2.5% over the year ending August 2026, with growth in traditionally higher-paid occupations accelerating to 2.6%.
What it means for pay decisions
For workers, AI exposure alone is a weak basis for salary negotiation. A data scientist with management duties, an entry-level analyst and an experienced legal specialist may use similar tools but face very different supply, demand and responsibility. A better comparison matches the same role, location, level and pay components. Employers should likewise avoid applying a blanket premium to any job carrying an AI label.
The finding pushes back on claims that AI is already squeezing pay in exposed professional work, but it does not settle questions of displacement, productivity or future wages. Watch whether measured wage surveys move in the same direction as advertised ranges.
Sources
- Indeed Hiring Lab (Jack Kennedy), “AI Exposure Isn’t Squeezing Advertised Pay in the US — It’s Boosting It,” September 17, 2026. https://hiringlab.indeed.com/2026/09/17/ai-exposure-isnt-squeezing-advertised-pay-in-the-us-its-boosting-it/
- Indeed Hiring Lab, “US Labor Market Snapshot: September 2026,” September 24, 2026. https://hiringlab.indeed.com/2026/09/24/us-labor-market-snapshot-september-2026/
Featured photo: A software developer at work in Zambia, from Wiki Loves Africa 2017. Representative image of high-AI-exposure office work; it is not connected to the Indeed data. Photo: Tsinkala, via Wikimedia Commons (CC BY-SA 4.0). Cropped from the original.
