Skip to content
Singulariki

AI exposure across work

One comparable scale · OpenAI GPTs + Felten AIOE

This is the same AI-exposure signal carried on every grouping page across Singulariki, gathered in one place so you can compare lenses. For each occupation we take two published studies — the OpenAI "GPTs are GPTs" human-rated share of tasks a large language model (with tools) could speed up by half or more, and the independent Felten/Raj/Seamans AI Occupational Exposure index — then average them across the occupations in each group. Because every lens uses the same two studies and the same percentile method, the Low / Moderate / High bands are comparable across all of them.

Most AI-exposed slices, any lens

  1. Media and Communication Workers Occupation groups High 68%
  2. Secretaries and Administrative Assistants Occupation groups High 67%
  3. Area, Ethnic, Cultural, Gender, and Group Studies Fields of study High 64%
  4. Mathematical Science Occupations Occupation groups High 64%
  5. Sales Representatives, Wholesale and Manufacturing Occupation groups High 62%
  6. Foreign Languages, Literatures, and Linguistics Fields of study High 60%
  7. English Language and Literature/Letters Fields of study High 59%
  8. Financial Specialists Occupation groups High 58%

Least AI-exposed slices, any lens

  1. Helpers, Construction Trades Occupation groups Low 0%
  2. Fishing and Hunting Workers Occupation groups Low 1%
  3. Construction Trades Workers Occupation groups Low 3%
  4. Assemblers and Fabricators Occupation groups Low 5%
  5. Grounds Maintenance Workers Occupation groups Low 5%
  6. Food Processing Workers Occupation groups Low 6%
  7. Extraction Workers Occupation groups Low 6%
  8. Vehicle and Mobile Equipment Mechanics, Installers, and Repairers Occupation groups Low 6%

By occupation browse all →

The grouping lenses above average this signal; here it is at full resolution — every individual occupation ranked on the same unified exposure index (950 occupations scored, 320 High / 315 Low). This is the SOC-native, directly-clickable companion to the international GenAI exposure gradient. Each occupation links to its full evidence page.

Most AI-exposed occupations

  1. News Analysts, Reporters, and Journalists High 100th pct
  2. Mathematicians High 100th pct
  3. Market Research Analysts and Marketing Specialists High 99th pct
  4. Search Marketing Strategists High 99th pct
  5. Data Scientists High 99th pct
  6. Business Intelligence Analysts High 99th pct
  7. Clinical Data Managers High 99th pct
  8. Operations Research Analysts High 99th pct
  9. Web Developers High 99th pct
  10. Statisticians High 99th pct
  11. Biostatisticians High 99th pct
  12. Economists High 99th pct

Least AI-exposed occupations

  1. Helpers--Roofers Low 0th pct
  2. Helpers--Painters, Paperhangers, Plasterers, and Stucco Masons Low 0th pct
  3. Foundry Mold and Coremakers Low 0th pct
  4. Orderlies Low 0th pct
  5. Pile Driver Operators Low 1st pct
  6. Helpers--Extraction Workers Low 1st pct
  7. Stonemasons Low 1st pct
  8. Rail-Track Laying and Maintenance Equipment Operators Low 1st pct
  9. Tire Builders Low 1st pct
  10. Roustabouts, Oil and Gas Low 1st pct
  11. Tire Repairers and Changers Low 1st pct
  12. Cement Masons and Concrete Finishers Low 1st pct

Percentile is each occupation's rank on the unified exposure index across all 950 scored occupations. Exposure measures task overlap with today's AI, not a prediction of automation.

By observed AI use how it's measured →

The lens above is potential — what AI could plausibly touch, from published task-overlap studies. This one is observed: Microsoft's "Working with AI" applicability score, the overlap between an occupation's work activities and activities seen in real Bing Copilot conversations (916 occupations scored, 307 High / 307 Low). It is activity overlap in assistant logs — not an adoption rate, automation, or replacement.

Highest observed AI applicability

  1. Sales Representatives of Services, Except Advertising, Insurance, Financial Services, and Travel High 100th pct
  2. Interpreters and Translators High 100th pct
  3. Writers and Authors High 100th pct
  4. Poets, Lyricists and Creative Writers High 100th pct
  5. Historians High 100th pct
  6. Customer Service Representatives High 99th pct
  7. Telemarketers High 99th pct
  8. Technical Writers High 99th pct
  9. News Analysts, Reporters, and Journalists High 99th pct
  10. Computer Numerically Controlled Tool Programmers High 99th pct
  11. Passenger Attendants High 99th pct
  12. Broadcast Announcers and Radio Disc Jockeys High 99th pct

Lowest observed AI applicability

  1. Dredge Operators Low 0th pct
  2. Bridge and Lock Tenders Low 0th pct
  3. Pile Driver Operators Low 0th pct
  4. Floor Sanders and Finishers Low 0th pct
  5. Foundry Mold and Coremakers Low 0th pct
  6. Rail-Track Laying and Maintenance Equipment Operators Low 0th pct
  7. Orderlies Low 0th pct
  8. Water and Wastewater Treatment Plant and System Operators Low 0th pct
  9. Motorboat Operators Low 1st pct
  10. Logging Equipment Operators Low 1st pct
  11. Roustabouts, Oil and Gas Low 1st pct
  12. Paving, Surfacing, and Tamping Equipment Operators Low 1st pct

Percentile is each occupation's rank on Microsoft's applicability score across all 916 scored occupations. Two studies, two methods: applicability is observed assistant activity; exposure is theoretical task overlap. They are not the same number and are not meant to be.

Where potential and observed disagree

The most revealing occupations are the ones where the two lenses pull apart. The gap is the difference between an occupation's potential exposure percentile and its observed applicability percentile — a structural fact about the data, not a claim about any worker.

High potential, lower observed use

AI could plausibly touch much of the work, but it shows up less in real assistant activity — frequently regulated, high-stakes, or verification-heavy work.

  1. Judicial Law Clerks exp 66th · obs 20th
  2. Labor Relations Specialists exp 62nd · obs 17th
  3. Title Examiners, Abstractors, and Searchers exp 61st · obs 17th
  4. Paralegals and Legal Assistants exp 56th · obs 13th
  5. Judges, Magistrate Judges, and Magistrates exp 62nd · obs 21st
  6. Natural Sciences Managers exp 74th · obs 34th
  7. Purchasing Managers exp 71st · obs 32nd
  8. Education Administrators, Postsecondary exp 67th · obs 29th

Lower potential, higher observed use

The task-overlap studies rank these lower, yet their activities surface more in assistant logs — often communication, information-gathering, and explanation that generalize across many jobs.

  1. Athletes and Sports Competitors exp 24th · obs 74th
  2. Models exp 47th · obs 96th
  3. Passenger Attendants exp 55th · obs 99th
  4. Waiters and Waitresses exp 44th · obs 83rd
  5. Riggers exp 22nd · obs 61st
  6. Amusement and Recreation Attendants exp 48th · obs 86th
  7. Cooks, Private Household exp 46th · obs 83rd
  8. Flight Attendants exp 40th · obs 77th

"exp" is the potential-exposure percentile; "obs" is the observed-applicability percentile, both across all occupations scored on each measure. Divergence is expected — the two studies ask different questions — and reading them together is more honest than trusting either alone.

Where AI exposure meets the job outlook

Two independent signals, placed on the same map: how exposed an occupation's tasks are to today's AI (the unified GPTs + AIOE index) and how fast the U.S. Bureau of Labor Statistics projects its employment to grow or shrink from 2024 to 2034. Neither causes the other — exposure is task overlap that usually means augmentation, and the BLS projection is built from demographics, demand and industry trends, not AI. Reading them together answers the question the headlines mangle: AI exposure is not the same thing as decline. The largest occupations in each quadrant are shown (918 occupations carry both signals).

Growing · more AI-exposed

Expanding fields where AI also overlaps the work — augmentation, not erasure.

  1. General and Operations Managers +4.4% · 60th exp
  2. Advanced Practice Psychiatric Nurses +4.9% · 54th exp
  3. Clinical Nurse Specialists +4.9% · 53rd exp
  4. Software Developers +15.8% · 91st exp
  5. Accountants and Auditors +4.6% · 93rd exp
  6. Sales Representatives, Wholesale and Manufacturing, Except Technical and Scientific Products +0.3% · 89th exp
  7. Sales Representatives of Services, Except Advertising, Insurance, Financial Services, and Travel +3.1% · 97th exp
  8. Business Operations Specialists, All Other +3.0% · 76th exp

Growing · less AI-exposed

Expanding and largely hands-on — demand growth with little task overlap today.

  1. Fast Food and Counter Workers +6.1% · 34th exp
  2. Baristas +6.1% · 36th exp
  3. Registered Nurses +4.9% · 47th exp
  4. Acute Care Nurses +4.9% · 46th exp
  5. Critical Care Nurses +4.9% · 46th exp
  6. Laborers and Freight, Stock, and Material Movers, Hand +1.5% · 4th exp
  7. Recycling and Reclamation Workers +1.5% · 4th exp
  8. Stockers and Order Fillers +8.5% · 29th exp

Declining · more AI-exposed

Shrinking and AI-overlapping — where pressure may compound, though BLS attributes the decline to broader forces.

  1. Retail Salespersons -0.5% · 62nd exp
  2. Cashiers -9.9% · 53rd exp
  3. Customer Service Representatives -5.5% · 94th exp
  4. Office Clerks, General -6.7% · 73rd exp
  5. Secretaries and Administrative Assistants, Except Legal, Medical, and Executive -1.6% · 83rd exp
  6. First-Line Supervisors of Office and Administrative Support Workers -0.3% · 80th exp
  7. Bookkeeping, Accounting, and Auditing Clerks -5.8% · 88th exp
  8. Elementary School Teachers, Except Special Education -2.0% · 59th exp

Declining · less AI-exposed

Shrinking for reasons other than AI — low task overlap, declining demand.

  1. Waiters and Waitresses -0.7% · 44th exp
  2. Food Preparation Workers -3.4% · 17th exp
  3. Shipping, Receiving, and Inventory Clerks -7.7% · 48th exp
  4. Cooks, Fast Food -13.5% · 30th exp
  5. Packers and Packagers, Hand -5.4% · 8th exp
  6. Childcare Workers -2.9% · 36th exp
  7. Nannies -2.9% · 35th exp
  8. Correctional Officers and Jailers -7.8% · 26th exp

Axes split at their natural midpoints: growth at 0% (BLS-projected expansion vs contraction) and exposure at the 50th percentile (more vs less exposed than the median occupation). "exp" is the occupation's exposure percentile; the percentage is its BLS projected 2024–2034 employment change. See the full outlook ranking and methodology.

Does AI exposure track pay?

The other half of the misread question: is AI exposure aimed at high-paid or low-paid work? The data refuses both easy answers. Crossing each occupation's median pay (BLS OEWS) with the same exposure index shows exposure tracking how textual, analytical and computational the tasks are — not the paycheck. It runs high in six-figure management and in lower-paid customer service and clerical work, and low in both low-wage hands-on service and high-wage bedside care. The largest occupations in each cell are shown (892 occupations carry both signals).

Exposure quadrant: AI task-overlap percentile vs Median pay AI task-overlap (horizontal) versus median pay (vertical), each a percentile across all scored occupations, for the 48 largest-employment U.S. occupations. Exposure runs high in well-paid cognitive work and in lower-paid clerical work alike — it tracks task structure, not the paycheck. Overlap measures shared tasks with AI, not automation. Lower overlap · higher pay Higher overlap · higher pay Higher overlap · lower pay Lower overlap · lower pay Landscaping and Groundskeeping Workers Construction Laborers Janitors and Cleaners, Except Maids and Housekeeping Cleaners Maintenance and Repair Workers, General Heavy and Tractor-Trailer Truck Drivers Light Truck Drivers Security Guards Fast Food and Counter Workers Acute Care Nurses First-Line Supervisors of Retail Sales Workers Elementary School Teachers, Except Special Education General and Operations Managers Retail Salespersons Secondary School Teachers, Except Special and Career/Technical Education Business Operations Specialists, All Other Online Merchants Bookkeeping, Accounting, and Auditing Clerks Human Resources Specialists Software Developers Customer Service Representatives AI task-overlap percentile → ↑ Median pay
The 48 largest-employment U.S. occupations, plotted by AI task-overlap percentile (horizontal) and median-pay percentile (vertical). The scatter shows no diagonal: exposure does not track pay. Overlap measures shared tasks with AI, not automation.

Higher-paid · more AI-exposed

Well-paid cognitive work — analysis, management, writing — where AI overlaps the tasks.

  1. General and Operations Managers $103k · 60th exp
  2. Advanced Practice Psychiatric Nurses $94k · 54th exp
  3. Clinical Nurse Specialists $94k · 53rd exp
  4. Software Developers $133k · 91st exp
  5. First-Line Supervisors of Office and Administrative Support Workers $66k · 80th exp
  6. Accountants and Auditors $82k · 93rd exp
  7. Sales Representatives, Wholesale and Manufacturing, Except Technical and Scientific Products $67k · 89th exp
  8. Sales Representatives of Services, Except Advertising, Insurance, Financial Services, and Travel $66k · 97th exp

Higher-paid · less AI-exposed

Well-paid for presence, hands or judgment — bedside, physical or in-person work.

  1. Registered Nurses $94k · 47th exp
  2. Acute Care Nurses $94k · 46th exp
  3. Critical Care Nurses $94k · 46th exp
  4. First-Line Supervisors of Construction Trades and Extraction Workers $79k · 42nd exp
  5. Solar Energy Installation Managers $79k · 48th exp
  6. Police and Sheriff's Patrol Officers $76k · 37th exp
  7. Customs and Border Protection Officers $76k · 41st exp
  8. Plumbers, Pipefitters, and Steamfitters $63k · 15th exp

Lower-paid · more AI-exposed

Lower-paid clerical, sales and support work that is heavily textual and procedural.

  1. Retail Salespersons $35k · 62nd exp
  2. Cashiers $31k · 53rd exp
  3. Customer Service Representatives $43k · 94th exp
  4. Office Clerks, General $44k · 73rd exp
  5. Secretaries and Administrative Assistants, Except Legal, Medical, and Executive $46k · 83rd exp
  6. Bookkeeping, Accounting, and Auditing Clerks $49k · 88th exp
  7. Elementary School Teachers, Except Special Education $62k · 59th exp
  8. First-Line Supervisors of Retail Sales Workers $47k · 55th exp

Lower-paid · less AI-exposed

Lower-paid hands-on service and manual work with little task overlap today.

  1. Fast Food and Counter Workers $30k · 34th exp
  2. Baristas $30k · 36th exp
  3. Laborers and Freight, Stock, and Material Movers, Hand $39k · 4th exp
  4. Recycling and Reclamation Workers $39k · 4th exp
  5. Stockers and Order Fillers $37k · 29th exp
  6. Waiters and Waitresses $34k · 44th exp
  7. Janitors and Cleaners, Except Maids and Housekeeping Cleaners $36k · 10th exp
  8. Heavy and Tractor-Trailer Truck Drivers $57k · 29th exp

Axes split at their medians: pay at the median occupation's wage ($63k) and exposure at the 50th percentile. "exp" is the occupation's exposure percentile; the dollar figure is its BLS median annual wage. Exposure is task overlap, not a wage forecast — see methodology.

The remote-work overlap

One of the least-noticed facts about this wave: the jobs that could move home in 2020 are largely the same jobs most exposed to AI now. Dingel and Neiman classified which occupations can be done entirely from home; crossing that with the exposure index shows the two pressures landing on the same desk-and-screen work. Of the remote-capable occupations here, 91% sit above the median on exposure; of the must-be-there occupations, 75% sit below it. Remote-capability is a structural flag (the task structure permits it), not a count of who actually works from home — and exposure is task overlap, not automation. 784 occupations carry both signals.

Remote-capable · more AI-exposed

The compounding case: work that can already be done from anywhere, and whose tasks overlap heavily with what AI does — clerical, administrative, analytical, financial.

  1. General and Operations Managers 60th exp
  2. Customer Service Representatives 94th exp
  3. Office Clerks, General 73rd exp
  4. Secretaries and Administrative Assistants, Except Legal, Medical, and Executive 83rd exp
  5. First-Line Supervisors of Office and Administrative Support Workers 80th exp
  6. Bookkeeping, Accounting, and Auditing Clerks 88th exp
  7. Accountants and Auditors 93rd exp
  8. Elementary School Teachers, Except Special Education 59th exp

Remote-capable · less AI-exposed

Can be done from home, but the tasks overlap little with current AI — small here, because remote-friendly work skews textual and AI-exposed.

  1. Childcare Workers 36th exp
  2. Nannies 35th exp
  3. Preschool Teachers, Except Special Education 46th exp
  4. Recreation Workers 47th exp
  5. Ushers, Lobby Attendants, and Ticket Takers 45th exp
  6. Detectives and Criminal Investigators 48th exp
  7. Speech-Language Pathology Assistants 46th exp
  8. Protective Service Workers, All Other 36th exp

Must be on-site · more AI-exposed

Presence is required, yet a real slice of the tasks — scheduling, records, communication, diagnosis — overlaps with AI. AI reshapes the desk-side of a hands-on job.

  1. Retail Salespersons 62nd exp
  2. Advanced Practice Psychiatric Nurses 54th exp
  3. Clinical Nurse Specialists 53rd exp
  4. Cashiers 53rd exp
  5. First-Line Supervisors of Retail Sales Workers 55th exp
  6. Receptionists and Information Clerks 76th exp
  7. First-Line Supervisors of Production and Operating Workers 59th exp
  8. Wind Energy Operations Managers 62nd exp

Must be on-site · less AI-exposed

The ground floor of the physical economy — moving, cleaning, driving, building, guarding — done in person with little current task overlap with AI.

  1. Registered Nurses 47th exp
  2. Acute Care Nurses 46th exp
  3. Critical Care Nurses 46th exp
  4. Laborers and Freight, Stock, and Material Movers, Hand 4th exp
  5. Recycling and Reclamation Workers 4th exp
  6. Waiters and Waitresses 44th exp
  7. Janitors and Cleaners, Except Maids and Housekeeping Cleaners 10th exp
  8. Heavy and Tractor-Trailer Truck Drivers 29th exp

Remote-capability: Dingel & Neiman, "How many jobs can be done at home?" (2020) — a feasibility classification, not observed remote rates. Exposure split at the 50th percentile; "exp" is the occupation's exposure percentile. This is a descriptive cross-tab, not a forecast of relocation or job loss — see methodology.

Then vs now: the 2013 reversal

In 2013, before large language models, Frey & Osborne estimated each occupation's "probability of computerisation." Their alarm pointed largely at routine manual and clerical work. A decade later, AI task-overlap exposure points somewhere strikingly different — at the cognitive, textual work 2013 rated safe. These are different measures (a whole-job automation guess then; task overlap now), which is exactly why the contrast is worth seeing. Crossing both percentiles recovers the reversal — and the places both eras still agree. 704 occupations carry both estimates.

2013 called safe · now AI-exposed

The reversal. Management, teaching, customer and business work Frey rated low-risk — now among the most overlapped by language models.

  1. General and Operations Managers 32nd '13 · 60th now
  2. Customer Service Representatives 49th '13 · 94th now
  3. First-Line Supervisors of Office and Administrative Support Workers 9th '13 · 80th now
  4. Elementary School Teachers, Except Special Education 3rd '13 · 59th now
  5. Business Operations Specialists, All Other 35th '13 · 76th now
  6. Business Continuity Planners 35th '13 · 87th now
  7. Sustainability Specialists 35th '13 · 87th now
  8. Online Merchants 35th '13 · 77th now

2013 called doomed · now low exposure

The over-call. Physical, in-person work 2013 expected to be computerised — but current AI overlaps its tasks little, because presence and hands still matter.

  1. Laborers and Freight, Stock, and Material Movers, Hand 71st '13 · 4th now
  2. Recycling and Reclamation Workers 71st '13 · 4th now
  3. Waiters and Waitresses 86th '13 · 44th now
  4. Janitors and Cleaners, Except Maids and Housekeeping Cleaners 56th '13 · 10th now
  5. Heavy and Tractor-Trailer Truck Drivers 64th '13 · 29th now
  6. Maintenance and Repair Workers, General 54th '13 · 20th now
  7. Cooks, Restaurant 91st '13 · 34th now
  8. Security Guards 69th '13 · 32nd now

Both eras agree · high

Routine cognitive work — clerical, sales, records — flagged then and exposed now. The one place the 2013 forecast and today's signal line up.

  1. Retail Salespersons 82nd '13 · 62nd now
  2. Cashiers 94th '13 · 53rd now
  3. Office Clerks, General 91st '13 · 73rd now
  4. Secretaries and Administrative Assistants, Except Legal, Medical, and Executive 91st '13 · 83rd now
  5. Bookkeeping, Accounting, and Auditing Clerks 97th '13 · 88th now
  6. Accountants and Auditors 86th '13 · 93rd now
  7. Sales Representatives, Wholesale and Manufacturing, Except Technical and Scientific Products 71st '13 · 89th now
  8. Receptionists and Information Clerks 91st '13 · 76th now

Both eras agree · low

Skilled hands-on trades and care — construction, electrical, medical assisting — that neither the 2013 estimate nor today's exposure index flags.

  1. First-Line Supervisors of Construction Trades and Extraction Workers 32nd '13 · 42nd now
  2. Solar Energy Installation Managers 32nd '13 · 48th now
  3. Medical Assistants 39th '13 · 37th now
  4. Electricians 31st '13 · 32nd now
  5. Police and Sheriff's Patrol Officers 28th '13 · 37th now
  6. Customs and Border Protection Officers 28th '13 · 41st now
  7. Licensed Practical and Licensed Vocational Nurses 24th '13 · 25th now
  8. Packers and Packagers, Hand 43rd '13 · 8th now

"'13" is the occupation's percentile on Frey & Osborne's 2013 probability of computerisation; "now" is its percentile on the current AI task-overlap exposure index. Both split at the 50th percentile. The two are different constructs and neither is an observed outcome — this is a contrast of estimates, not a forecast. See methodology.

Does more school protect you?

The intuition that education is a shield against automation runs backwards here. Joining each occupation's typical entry education (BLS) to the exposure index and weighting by employment, exposure rises with required schooling up through a bachelor's degree — then eases at the master's and doctoral tier, where hands-on professions like physicians, dentists and surgeons mix in. The work that needs no formal credential — physical, in-person — is the least exposed. Exposure is task overlap with AI, not automation; "entry education" is the BLS category for the occupation, not a statement about any person in it.

  1. No credential 28th pct

    Retail Salespersons · Fast Food and Counter Workers · Baristas

  2. High school 52nd pct

    Customer Service Representatives · Office Clerks, General · Secretaries and Administrative Assistants, Except Legal, Medical, and Executive

  3. Nondegree award 28th pct

    Heavy and Tractor-Trailer Truck Drivers · Nursing Assistants · Medical Assistants

  4. Associate's 50th pct

    Preschool Teachers, Except Special Education · Paralegals and Legal Assistants · Radiologic Technologists and Technicians

  5. Bachelor's 71st pct

    General and Operations Managers · Registered Nurses · Acute Care Nurses

  6. Master's 62nd pct

    Educational, Guidance, and Career Counselors and Advisors · Education Administrators, Kindergarten through Secondary · Nurse Practitioners

  7. Doctoral / pro 60th pct

    Lawyers · Pharmacists · Physicians, All Other

Each bar is the employment-weighted mean exposure percentile of the occupations whose typical entry education is that tier (tiers with fewer than 10 occupations omitted). Examples are the largest occupations by employment in each tier. See methodology.

By industry browse all →

Each industry's occupations weighted by employment, then by the unified exposure index — so this reads the real labor structure of the sector, not a flat occupation average. This is the employment-weighted lens; the grouping tiers below use unweighted occupation means. The split is the cleanest refutation of both lazy takes: AI does not touch "everything," and it does not skip "blue collar" — it tracks how textual, analytical, and computational the work actually is.

Most AI-exposed industries

  1. Insurance Agencies and Brokerages High 99th pct
  2. Finance and Insurance High 98th pct
  3. Direct Health and Medical Insurance Carriers High 96th pct
  4. Research and Development in the Social Sciences and Humanities High 95th pct
  5. Radio Broadcasting Stations High 93rd pct
  6. Newspaper Publishers High 92nd pct
  7. Management of Companies and Enterprises High 90th pct
  8. Information High 89th pct
  9. Professional, Scientific, and Technical Services High 87th pct
  10. Television Broadcasting Stations High 86th pct
  11. Engineering Services High 84th pct
  12. Labor Unions and Similar Labor Organizations High 83rd pct

Least AI-exposed industries

  1. Landscaping Services Low 1st pct
  2. Agriculture, Forestry, Fishing and Hunting Low 2nd pct
  3. Poured Concrete Foundation and Structure Contractors Low 4th pct
  4. Masonry Contractors Low 5th pct
  5. Painting and Wall Covering Contractors Low 7th pct
  6. Drywall and Insulation Contractors Low 8th pct
  7. Roofing Contractors Low 10th pct
  8. Geothermal Electric Power Generation Low 11th pct
  9. Ambulance Services Low 13th pct
  10. Mining, Quarrying, and Oil and Gas Extraction Low 14th pct
  11. Power and Communication Line and Related Structures Construction Low 16th pct
  12. Transportation and Warehousing Low 17th pct

Employment-weighted mean of the unified exposure index across each industry's occupations that carry a score (BLS national industry-occupation matrix, May 2024). Percentile is across all 67 industries. Exposure = task overlap with today's AI, not automation.

Career clusters browse all →

The 14 broad families of related careers schools and workforce systems organize around.

Hottest in this lens AI exposure Avg task overlap
Digital Technology High 58%
Financial Services High 55%
Marketing & Sales High 50%
Management & Entrepreneurship High 49%
Education High 45%
Public Service & Safety Moderate 44%
coolest: Construction Low 13%

Education levels browse all →

The 8 typical entry-level education tiers BLS assigns to occupations.

Hottest in this lens AI exposure Avg task overlap
Bachelor's degree High 47%
Master's degree High 47%
Doctoral or professional degree High 40%
Some college, no degree Moderate 39%
Associate's degree Moderate 32%
High school diploma or equivalent Low 20%
coolest: No formal educational credential Low 11%

Preparation zones browse all →

O*NET job zones — how much education, experience and training a job needs.

Hottest in this lens AI exposure Avg task overlap
Job Zone Four: Considerable Preparation Needed High 46%
Job Zone Five: Extensive Preparation Needed Moderate 43%
Job Zone Three: Medium Preparation Needed Moderate 28%
Job Zone 1-2: Very Little to Some Preparation Needed Low 15%

Fields of study browse all →

CIP-2020 instructional-program families — what you study, mapped to where it leads.

Hottest in this lens AI exposure Avg task overlap
Area, Ethnic, Cultural, Gender, and Group Studies High 64%
Foreign Languages, Literatures, and Linguistics High 60%
English Language and Literature/Letters High 59%
Mathematics and Statistics High 57%
Computer and Information Sciences and Support Services High 53%
Communication, Journalism, and Related Programs High 52%
coolest: Precision Production Low 8%

Job families browse all →

The 22 major SOC occupation groups — the top level of the federal job taxonomy.

Hottest in this lens AI exposure Avg task overlap
Computer and Mathematical Occupations High 57%
Business and Financial Operations Occupations High 52%
Sales and Related Occupations High 49%
Life, Physical, and Social Science Occupations High 49%
Office and Administrative Support Occupations High 49%
Management Occupations High 43%
coolest: Construction and Extraction Occupations Low 6%

Occupation groups browse all →

The finer SOC minor groups nested inside the job families.

Hottest in this lens AI exposure Avg task overlap
Media and Communication Workers High 68%
Secretaries and Administrative Assistants High 67%
Mathematical Science Occupations High 64%
Sales Representatives, Wholesale and Manufacturing High 62%
Financial Specialists High 58%
Sales Representatives, Services High 57%
coolest: Helpers, Construction Trades Low 0%

Group averages are unweighted across member occupations that carry a published exposure score; the industry lens is employment-weighted. Bands: Low / Moderate / High by cross-group percentile within each tier, comparable across tiers by construction.

Sources for this page

Every figure above traces to a named public dataset and the exact release below — not hand-written opinion. See the full methodology for what each measure does and does not mean.