AI-exposure ranking

The 50 safest occupations from AI automation

Ranked by AI-exposure score, O*NET task analysis of more than 18,000 occupational tasks across 832 occupations. According to U.S. Bureau of Labor Statistics Employment Projections 2024–2034 (O*NET 30.0, as of May 2024), the least-exposed work is hands-on and hard to automate. See our methodology for how each score is computed.

Safest role
Helpers--electricians
Its AI exposure
5%
Top-10 average
12.1%
Also growing
36/50
The short version

According to U.S. Bureau of Labor Statistics Employment Projections 2024–2034 and O*NET task analysis of more than 18,000 occupational tasks across 832 occupations, the least AI-exposed work is hands-on and unpredictable - skilled trades, hands-on healthcare, and physical service roles. Helpers--electricians tops the list at just 5% exposure, and 36 of these 50 occupations are also projected to add jobs through 2034, resilience and demand in the same roles.

Scores reflect O*NET 30.0 task data (as of BLS May 2024); see our methodology for how each occupation's AI-exposure score is computed.

The full ranking

# Occupation
1 Helpers--electricians
2 Physicians, all other
3 Manufactured building and mobile home installers
4 Pediatricians, general
5 Electricians
6 Farm equipment mechanics and service technicians
7 Aircraft mechanics and service technicians
8 Landscaping and groundskeeping workers
9 Stonemasons
10 Drywall and ceiling tile installers
11 Recreational vehicle service technicians
12 Orderlies
13 Cement masons and concrete finishers
14 Terrazzo workers and finishers
15 Construction laborers
16 Excavating and loading machine and dragline operators, surface mining
17 Millwrights
18 Helpers--installation, maintenance, and repair workers
19 Agricultural equipment operators
20 Telecommunications equipment installers and repairers, except line installers
21 Structural metal fabricators and fitters
22 Healthcare diagnosing or treating practitioners, all other
23 Farmworkers and laborers, crop, nursery, and greenhouse
24 Paving, surfacing, and tamping equipment operators
25 Roofers
26 Helpers--brickmasons, blockmasons, stonemasons, and tile and marble setters
27 Refractory materials repairers, except brickmasons
28 Foundry mold and coremakers
29 Healthcare social workers
30 Mental health and substance abuse social workers
31 Glaziers
32 Mechanical door repairers
33 Electrical power-line installers and repairers
34 Wind turbine service technicians
35 Engine and other machine assemblers
36 Clergy
37 Prosthodontists
38 General internal medicine physicians
39 Orthopedic surgeons, except pediatric
40 Farmworkers, farm, ranch, and aquacultural animals
41 Psychiatrists
42 Pediatric surgeons
43 Occupational therapy aides
44 Animal caretakers
45 Lighting technicians
46 Bicycle repairers
47 Musical instrument repairers and tuners
48 Maintenance and repair workers, general
49 Cleaners of vehicles and equipment
50 Athletes and sports competitors

Source: O*NET task data (U.S. Department of Labor) for AI-exposure analysis by PlainWorkforce; BLS Employment Projections 2024–2034 for growth and wages. Ranked across 832 tracked occupations.

How this ranking is built

Each occupation's AI-exposure score derives from O*NET 30.0 task data: every task is rated for automation potential, with routine cognitive work (data entry, monitoring) weighted high and non-automatable work (manual dexterity, social perceptiveness, judgment in unstructured settings) weighted near zero. Task scores are aggregated by importance and level, then normalized to a 0–100% occupation score. Lower scores rank higher here. Employment growth and wage figures come from BLS Employment Projections 2024–2034.

Caveats and limitations

Scores estimate current task automatability and may shift as AI capabilities advance. The analysis focuses on replacement risk, not augmentation; it omits geographic and demographic variation beyond national aggregates; and BLS projections carry their own uncertainty intervals. Use it as a directional guide, not a guarantee.

Every figure on PlainWorkforce is rendered directly from BLS/O*NET source data, no number is typed in by an editor. See our editorial standards & corrections policy, the methodology behind these figures, or report a data error. Data current as of BLS Employment Projections 2024-2034.