METHODOLOGY
How we calculate automation risk — and why you should care.
Our Data Sources
We don't make this up. Every prediction is grounded in peer-reviewed research and government labor data. Here's what powers the numbers:
Frey & Osborne (2013, updated 2017)
"The Future of Employment: How Susceptible Are Jobs to Computerisation?" — the landmark Oxford study that scored 702 occupations from 0% to 99% automation probability.
McKinsey Global Institute
"Jobs Lost, Jobs Gained" (2017) and "The Future of Work After COVID-19" (2021). McKinsey predicts that 30% of tasks in 60% of all occupations could be automated by 2030.
World Economic Forum
"The Future of Jobs Report" (2020, 2023, 2025). The WEF flags 85 million jobs at risk globally by 2025 but projects 97 million new ones emerging.
U.S. Bureau of Labor Statistics (BLS)
Real employment counts, median wages, and occupation growth/decline projections from the BLS Occupational Employment and Wage Statistics program.
O*NET (Department of Labor)
Detailed task and skill breakdowns for each occupation — manual dexterity, creativity, social perceptiveness, and more.
80,000 Hours — Skills AI Makes More Valuable (2025)
Identifies 6 resilient skill categories that increase in value as AI progresses (AI application, personal effectiveness, leadership, communication & taste, government & policy, complex physical skills) and 4 vulnerable categories. Key insight: automation of a task doesn't always decrease wages — ATMs initially increased bank teller employment.
Epoch AI — GATE Model
Integrated economic model of automation unfolding task by task. Default assumptions show wages initially increasing ~10x up to 2037, then collapsing as final bottlenecks are removed. With 1% of jobs legally mandated human-in-the-loop, wages increase indefinitely.
How We Calculate Risk
Base Probability
Start with the Frey & Osborne score (0-99%) for the closest matching occupation.
Task Decomposition
Break the job into individual tasks using O*NET data. Score each task for routine-ness, creativity, social skills, and physical requirements.
Timeline Estimation
Cross-reference with McKinsey's task-level automation forecasts and current AI capability trajectories.
Skill Value Shift
Apply the 80,000 Hours framework: classify each skill as resilient (gains value with AI) or vulnerable (loses value). Check for the ATM effect — whether automation initially increases demand.
Pivot Generation
Identify adjacent occupations that build on your resilient skills while avoiding your vulnerable ones.
Risk Categories
These jobs are actively being replaced NOW. Most tasks are routine and digital. Timeline: 1-5 years.
Major disruption incoming. Core tasks being automated but some human elements remain. Timeline: 3-10 years.
AI becomes a co-pilot, not a replacement. Your job changes significantly but doesn't disappear. Timeline: 5-15 years.
Requires physical presence, deep empathy, creative judgment, or complex human interaction. Timeline: 15+ years.
The Key Insight
"No major job category faces majority automation. 61% of US workers — 142.8 million people — are in the 'AI Co-Pilot Zone' where AI assists rather than replaces their work."
— WILLAI 2025 AI Job Impact Report, analyzing 922 occupations and 57,326 work tasks
This means the question isn't "will my job exist?" — it's "what will my job look like?" The people who thrive will be the ones who learn to work WITH AI tools, not compete against them.
Betting Market Odds
The betting market uses an odds formula that factors in both the academic automation probability and your chosen time horizon:
Shorter time horizons have higher base multipliers (5x for 2-year, 3x for 5-year, etc.). Betting that a LOW-risk job will be automated SOON gives enormous odds — because it's very unlikely. Betting "Never" on a HIGH-risk job is the contrarian play and also pays well.
All bets use play money (₿). You start with ₿10,000. It's about being right, not being rich.