Top-10 Most AI-Exposed Occupations

Data

Per massenkoff-mccrory-labor-market-impacts-2026 Figure 3. Observed exposure measure (see observed-exposure-measure), Q3–Q4 2025 Claude work-context usage:

Rank	Occupation	Observed exposure	Leading automated task
1	Computer programmers	74.5%	Write, update, and maintain software programs
2	Customer service representatives	70.1%	Confer with customers to provide info, take orders, handle complaints
3	Data entry keyers	67.1%	Read source documents and enter data into systems
4	Medical record specialists	66.7%	Compile, abstract, and code patient data
5	Market research analysts and marketing specialists	64.8%	Prepare reports of findings, illustrating data graphically and translating complex findings into written text
6	Sales representatives, wholesale and manufacturing (except technical and scientific products)	62.8%	Contact customers to demonstrate products and solicit orders
7	Financial and investment analysts	57.2%	Inform investment decisions by analyzing financial information to forecast business, industry, or economic conditions
8	Software quality assurance analysts and testers	51.9%	Modify software to correct errors or improve performance
9	Information security analysts	48.6%	Perform risk assessments and test data processing security
10	Computer user support specialists	46.8%	Answer user inquiries regarding computer software or hardware operation to resolve problems

Zero exposure. 30% of US workers have zero observed exposure — cooks, motorcycle mechanics, lifeguards, bartenders, dishwashers, dressing room attendants.
Software developers (distinct from “computer programmers”) sit at ~30% exposure but with very high BLS-projected growth — the value-stack-top outlier in ai-exposure-vs-bls-growth.
Cashiers sit near 10% exposure with sharply negative projected growth (−5 to −7%).
Customer service reps: highest exposure with negative BLS growth (~−2.5%) — the canonical example of an exposure-growth alignment.

Coding dominates. Three of the top ten (#1 programmers, #8 SQA, #10 support) are software-adjacent. Reinforces middle-layer-defensibility story: Cursor, Windsurf, Copilot are deployed against this exposure.
Knowledge work is over-represented. Analysts (financial, market research, information security) appear three times; medical and customer-service white-collar roles three times.
No purely manual labor in the top 10. Lowest exposure cohort is uniformly physical-presence work (food service, repair, hospitality).
The 50% threshold is meaningful. All ten occupations have observed exposure above 46%. The next tier (accountants, lawyers, registered nurses) sit in the 25–40% range per the radar plot in Figure 2 of the source.

Labor-side anchor for the H2_u-curve-of-value top. The customers of L4–L6 application-layer firms are this list. Cursor’s user base is occupation #1; LangChain/agent platforms are tooled for #5–7 analyst work; Claude API is the substrate for #2 customer service.
Empirical correlate of exposed-worker-demographics. The top-10 list, weighted by employment, produces the demographic profile in the worker characteristics table.
Reframes the bear-case incidence. The Acemoglu (2024) prediction emphasised low-education clerical work; this list is dominated by college-educated knowledge work. Software QA and information security in particular are mid-six-figure roles.

massenkoff-mccrory-labor-market-impacts-2026 — source.
observed-exposure-measure — measurement methodology.
exposed-worker-demographics — who these workers are.
ai-exposure-vs-bls-growth — projected growth for these occupations.
H2_u-curve-of-value — the value-stack consequence.
capital-labor-divergence — surplus-distribution story.
middle-layer-defensibility — what L4–L6 captures by serving these workers.
ai-young-worker-hiring-slowdown — entry-level cohort in these occupations is the canary signal.