Jobpocalypse AI OVERLAP INDEX
v1.0 - April 2026 342 occupations 2024-34 projections (BLS August 2025 release)

The AI Overlap Index for 342 US occupations.

A research-grade scoring of how exposed each US occupation is to current-generation AI. Built on the BLS 2024-2034 employment projections released August 28, 2025, which forecast 5.2 million new jobs over the decade. Claude-scored task automation plus four weighted signals, mapped against ten years of projected employment growth.

Occupations
342
scored
Mean AOI
44.8
0-100
Highly/Mostly Exposed
69
20% of roles
Insulated
33
most insulated
Categories
25
BLS groupings

The map

Tile area scales with AOI. Color encodes the exposure tier. Click a category block to zoom; click a tile to open the profile.

data: aoi.json · 342 tiles · d3-treemap
Highly Exposed Mostly Exposed Partially Exposed Selectively Exposed Insulated

Exposure tiers

Five cut-points: 70+, 55-70, 40-55, 30-40, below 30.

Highly Exposed 7
Tasks are text, numbers, code, or routine decisions. Productivity tools already cover the bulk of the work.
Mostly Exposed 62
Most of the workflow is automatable. Human judgment remains for exceptions, clients, or ambiguity.
Partially Exposed 150
Clear pressure on routine tasks. Composition of the role will shift within the decade.
Selectively Exposed 90
Physical, social, or oversight-heavy work that AI augments rather than replaces.
Insulated 33
Embodied skill, frontline presence, or deep institutional judgment. Most insulated.

Most exposed to AI displacement

top 10 by AI Overlap Index

1
Medical transcriptionists Healthcare
80.1
2
Bookkeeping, accounting, and auditing clerks Office And Administrative Support
74.6
3
Customer service representatives Office And Administrative Support
73.2
4
Financial clerks Office And Administrative Support
72.8
5
Travel agents Sales
72.5
6
Material recording clerks Office And Administrative Support
70.6
7
Desktop publishers Office And Administrative Support
70.2
8
Secretaries and administrative assistants Office And Administrative Support
69.8
9
Insurance sales agents Sales
69.7
10
Bill and account collectors Office And Administrative Support
69.2

Most resilient

bottom 10 by AI Overlap Index

1
Construction laborers and helpers Construction And Extraction
18.9
2
Insulation workers Construction And Extraction
22.2
3
Cooks Food Preparation And Serving
23.7
4
Military careers Military
24.2
5
Surgical assistants and technologists Healthcare
24.2
6
Grounds maintenance workers Building And Grounds Cleaning
24.7
7
Small engine mechanics Installation Maintenance And Repair
24.8
8
Drywall installers, ceiling tile installers, and tapers Construction And Extraction
26.0
9
Veterinary assistants and laboratory animal caretakers Healthcare
26.3
10
Fire inspectors Protective Service
26.7
See all 342 occupations →

Categories, ranked by mean AOI

Aggregates across the BLS occupational groupings. Bar shows mean AOI relative to the top category.

Office And Administrative Support
64.4
n=13
Sales
59.9
n=10
Math
56.8
n=4
Computer And Information Technology
56.0
n=10
Business And Financial
55.1
n=24
Media And Communication
54.2
n=10
Legal
52.2
n=5
Arts And Design
52.0
n=9
Architecture And Engineering
48.8
n=29
Management
47.0
n=26
Life Physical And Social Science
46.9
n=29
Education Training And Library
44.8
n=14
Production
43.0
n=16
Community And Social Service
42.1
n=9
Transportation And Material Moving
40.0
n=11
Personal Care And Service
39.1
n=11
Food Preparation And Serving
38.0
n=6
Healthcare
37.8
n=49
Entertainment And Sports
37.6
n=8
Protective Service
35.4
n=6
Farming Fishing And Forestry
34.5
n=4
Installation Maintenance And Repair
32.6
n=15
Building And Grounds Cleaning
31.9
n=3
Construction And Extraction
30.2
n=20
Military
24.2
n=1

How the AOI is built

The AI Overlap Index aggregates four weighted signals: task automation impact from per-task LLM scoring (60%), automation potential from Karpathy/BLS Digital AI Exposure (10%), market pressure from BLS outlook (15%), and entry-barrier erosion from BLS entry education (15%). Higher means more of the occupation's daily work is within reach of current AI systems. It is a directional descriptor, not a probability of job loss.

Every score is traceable to a task-level LLM evaluation, a BLS field, or a cited external prior. See the methodology for formulas, prompts, and weights.

Read the methodology → About the project
The four signals
60%
Task Automation Impact
Average LLM-rated automation score across the role's tasks, weighted by task importance (Core 3x, Important 2x, Supporting 1x).
10%
Automation Potential
Karpathy/BLS Digital AI Exposure (0-10 scale rescaled to 0-100), used directly.
15%
Market Pressure
BLS 2024-2034 outlook translated to a pressure score. Shrinking roles score high; growing roles score low.
15%
Entry Barrier Erosion
How far AI is lowering the threshold to enter the field, indexed to BLS typical entry-level education.