AVGPAY CROSS-DATASET RESEARCH

Which highly AI-exposed jobs are still projected to grow?

Exposure to AI and demand for workers are not opposites. This research looks for occupations with relatively high modeled task exposure where BLS nonetheless projects meaningful employment growth over its published decade.

The answer from the data

Among 87 matched occupations, Nurse Practitioner has one of the strongest projected changes at +41.0%, despite modeled exposure of 40/100. That contrast is the central finding: the direction of expected demand cannot be deduced from task exposure alone.

87Matching occupations
683Occupations screened
May 2025Published wage reference

These are screening results. The table shows the leading 15 matching occupations according to the stated ordering, not an unrestricted ranking of every U.S. job.

Why this comparison is useful

AI tools can change how a job is performed without removing demand for the service itself. More efficient workflows, expanding industries and new responsibilities can coexist with exposure. A cross-dataset view prevents the false inference that high exposure equals certain decline.

This analysis joins occupation codes across independently collected sources. Median wages are national occupational summaries, AI scores reflect the modeled share of tasks potentially accelerated under historic assumptions, and employment forecasts describe the BLS outlook for the entire occupation. A shared occupation label does not imply a uniform experience for everyone in that profession.

The occupations that match

Open any occupation for the full wage range, work description, skills, career paths and any evidence-backed salary scenarios. The separate AI link leads to its task-exposure research.

OccupationNational medianAI task exposureBLS projected growthDistinctive result
Nurse PractitionerAI research →$132,30040/100+41.0%+41.0%
Data ScientistAI research →$120,23064/100+34.6%+34.6%
Healthcare AdministratorAI research →$123,86040/100+24.2%+24.2%
Computer Research ScientistAI research →$140,30057/100+21.8%+21.8%
Cybersecurity AnalystAI research →$129,18060/100+21.0%+21.0%
Safety TechnicianAI research →$61,56041/100+19.4%+19.4%
EpidemiologistAI research →$87,22050/100+18.7%+18.7%
Health Sciences ProfessorAI research →$107,31043/100+17.9%+17.9%
LogisticianAI research →$82,32056/100+17.6%+17.6%
Speech TherapistAI research →$97,87042/100+16.6%+16.6%
Health Information SpecialistAI research →$68,02059/100+15.9%+15.9%
IT ManagerAI research →$175,14055/100+15.8%+15.8%
Medical ScientistAI research →$103,41055/100+12.6%+12.6%
Industrial EngineerAI research →$102,44054/100+12.4%+12.4%
BiochemistAI research →$127,41045/100+12.3%+12.3%

Sort order is fixed to the primary question's metric. The table is truncated at fifteen rows for readability; 87 occupations satisfy the criteria.

Exactly how the occupations were selected

Modeled task exposure at least 40 out of 100, projected ten-year BLS employment growth of at least 5%, a direct study match, and at least 10,000 workers. Ranked by projected growth, highest first.

We also exclude generic occupation titles containing “all other” to avoid ambiguous broad categories. Results use only locally verified, published datasets and rerun with each validated data refresh and build. No search-volume estimate is used to decide which jobs match.

What this analysis does not prove

BLS employment projections include assumptions about technological change, so they are not independent evidence that AI causes growth. Projections are scenarios for the economy, not guaranteed hiring, and task-exposure scores describe possible acceleration rather than adoption.

Occupational data can lag changing hiring practices. Differences in industry, region, hours, occupational mix, sampling and classification limit what can be inferred about a particular offer or individual worker. No causal relationship between AI scores and changes in employment or wages is established by joining these datasets.

How to use the findings

Examine whether demand comes from an expanding market, credential barriers, service needs or other drivers. Then ask which tasks actually change and how job requirements could shift.

Pick two or three realistic occupations from the table. Open the salary profiles to compare percentile wages, location and typical entry requirements, then review the AI research for task-level questions—not just a score. If the target requires a license or substantial retraining, model those costs before deciding.

Compare two occupations side by side → · Explore career payback →

Sources, time periods and reproducibility

Wages: U.S. BLS Occupational Employment and Wage Statistics, May 2025. Employment outlook: 2025–2035 BLS Employment Projections. AI exposure: historical GPT task assessment (2023 assumptions), direct matches where specified. Source vintages differ, so this analysis is descriptive rather than a combined statistical forecast.

Detailed AvgPay data methodology and sources → · Workforce-wide AI research report →