Futureproof
Menu

United States · release 2026-q4.1

Which jobs AI can already do most of

778 United States occupations, ranked by how much of the work inside them these tools can already do. The most exposed and the least exposed are on the same screen on purpose.

A high score does not mean the job disappears. It means a large share of the tasks inside it are now things software can do, which is a statement about the work, not about the person or about demand. Every job here has its own page showing exactly which tasks drove the number.

Most and least exposed occupations

Most exposed

The largest share of this work is already inside what these tools can do.

  1. 1Dispatchers, Except Police, Fire, and Ambulance96%of its work already doableshifting to AI
  2. 2Insurance Claims and Policy Processing Clerks94%of its work already doableshifting to AI
  3. 3Payroll and Timekeeping Clerks93%of its work already doableshifting to AI
  4. 4Web Developers89%of its work already doableshifting to AI
  5. 5Computer Numerically Controlled Tool Programmers87%of its work already doableshifting to AI
  6. 6Mathematical Science Occupations, All Other85%of its work already doableshifting to AI
  7. 7Data Scientists84%of its work already doableshifting to AI
  8. 8Cartographers and Photogrammetrists84%of its work already doableshifting to AI
  9. 9Database Architects83%of its work already doableshifting to AI
  10. 10Database Administrators82%of its work already doableshifting to AI
  11. 11Desktop Publishers81%of its work already doableshifting to AI
  12. 12Proofreaders and Copy Markers81%of its work already doableshifting to AI
  13. 13Statistical Assistants80%of its work already doableshifting to AI
  14. 14Correspondence Clerks79%of its work already doableshifting to AI
  15. 15Medical Transcriptionists79%of its work already doableshifting to AI
  16. 16Software Quality Assurance Analysts and Testers78%of its work already doableshifting to AI
  17. 17Credit Analysts78%of its work already doableshifting to AI
  18. 18Special Effects Artists and Animators78%of its work already doableshifting to AI
  19. 19Medical Records Specialists76%of its work already doableshifting to AI
  20. 20Credit Authorizers, Checkers, and Clerks76%of its work already doableshifting to AI
  21. 21Travel Agents75%of its work already doableshifting to AI
  22. 22Logisticians74%of its work already doableshifting to AI
  23. 23Cargo and Freight Agents73%of its work already doableshifting to AI
  24. 24Insurance Underwriters72%of its work already doableshifting to AI
  25. 25Graphic Designers72%of its work already doableshifting to AI

Least exposed

Physical presence, licensed accountability and real-time judgement are what hold these up.

  1. 588Woodworking Machine Setters, Operators, and Tenders, Except Sawing0%of its work already doablestaying human
  2. Wind Turbine Service Technicians0%of its work already doablestaying human
  3. Wellhead Pumpers0%of its work already doablestaying human
  4. Welding, Soldering, and Brazing Machine Setters, Operators, and Tenders0%of its work already doablestaying human
  5. Welders, Cutters, Solderers, and Brazers0%of its work already doablestaying human
  6. Ushers, Lobby Attendants, and Ticket Takers0%of its work already doablestaying human
  7. Tree Trimmers and Pruners0%of its work already doablestaying human
  8. Tool Grinders, Filers, and Sharpeners0%of its work already doablestaying human
  9. Tire Builders0%of its work already doablestaying human
  10. Tile and Stone Setters0%of its work already doablestaying human
  11. Terrazzo Workers and Finishers0%of its work already doablestaying human
  12. Teaching Assistants, Except Postsecondary0%of its work already doablestaying human
  13. Taxi Drivers0%of its work already doablestaying human
  14. Tapers0%of its work already doablestaying human
  15. Tailors, Dressmakers, and Custom Sewers0%of its work already doablestaying human
  16. Surgical Assistants0%of its work already doablestaying human
  17. Subway and Streetcar Operators0%of its work already doablestaying human
  18. Substitute Teachers, Short-Term0%of its work already doablestaying human
  19. Structural Iron and Steel Workers0%of its work already doablestaying human
  20. Stonemasons0%of its work already doablestaying human
  21. Slaughterers and Meat Packers0%of its work already doablestaying human
  22. Shuttle Drivers and Chauffeurs0%of its work already doablestaying human
  23. Shoe Machine Operators and Tenders0%of its work already doablestaying human
  24. Shoe and Leather Workers and Repairers0%of its work already doablestaying human
  25. Sheet Metal Workers0%of its work already doablestaying human

Showing 25 at each end of 778. Browse all United States occupations to see the middle, or search for your own job.

Read the range, not the position

Each score is published with the range it could reasonably sit in, and neighbouring rows usually overlap almost completely. So being 4th rather than 9th is not a finding, and we are not going to dress it up as one. What the ends of this table support is the broad shape: clerical, coordination and document work at one end, hands-on and licensed work at the other.

The ranks are here because they make a long table readable, not because position 12 means something position 13 does not.

How the score is calculated →

By field of work

Whole fields, not single jobs

The official major groups of the US classification, ordered by the average score of the jobs inside them. The spread matters more than the average: a field whose jobs run from 12 to 79 does not really have an average, and the column on the right shows you when that is the case.

United States major occupation groups by mean AI exposure score
FieldJobsAverageSpread across the field
Computer and Mathematicalmost exposed: Mathematical Science Occupations, All Other216448 to 78
Business and Financial Operationsmost exposed: Credit Analysts325520 to 70
Office and Administrative Supportmost exposed: Proofreaders and Copy Markers50547 to 80
Managementmost exposed: Computer and Information Systems Managers364633 to 57
Sales and Relatedmost exposed: Travel Agents214319 to 76
Legalmost exposed: Title Examiners, Abstractors, and Searchers74220 to 65
Architecture and Engineeringmost exposed: Cartographers and Photogrammetrists354015 to 69
Life, Physical, and Social Sciencemost exposed: Social Science Research Assistants473915 to 63
Arts, Design, Entertainment, Sports, and Mediamost exposed: Technical Writers36364 to 69
Educational Instruction and Librarymost exposed: Tutors60359 to 50
Community and Social Servicemost exposed: Health Education Specialists133119 to 55
Healthcare Practitioners and Technicalmost exposed: Medical Records Specialists69233 to 63
Protective Servicemost exposed: Gambling Surveillance Officers and Gambling Investigators23204 to 41
Healthcare Supportmost exposed: Medical Transcriptionists16181 to 71
Personal Care and Servicemost exposed: First-Line Supervisors of Personal Service Workers27185 to 36
Transportation and Material Movingmost exposed: Air Traffic Controllers46141 to 40
Building and Grounds Cleaning and Maintenancemost exposed: First-Line Supervisors of Housekeeping and Janitorial Workers8131 to 35
Food Preparation and Serving Relatedmost exposed: Chefs and Head Cooks15120 to 35
Installation, Maintenance, and Repairmost exposed: First-Line Supervisors of Mechanics, Installers, and Repairers50121 to 34
Productionmost exposed: Computer Numerically Controlled Tool Programmers99110 to 71
Farming, Fishing, and Forestrymost exposed: First-Line Supervisors of Farming, Fishing, and Forestry Workers11101 to 26
Construction and Extractionmost exposed: Construction and Building Inspectors5660 to 44
What this ranking leaves out (99 occupations)
  • 47 published occupations have no official task list at all, so there is nothing to measure and never will be. These are the “All Other” residual groups the classification uses to catch everything that does not fit elsewhere.
  • 52 are scored from fewer than 5 task statements. The number is real, but it is too thin a base to summarise a whole job, and those rows would otherwise sit at the extremes of this table purely because of how few tasks they have.
  • 0 have task statements that are not yet scored.

All of them still have pages, and every one says which of these applies to it.