The FAIR Framework
Futurist-Informed AI Risk — Developed by James Perkins
The FAIR Framework (Futurist-Informed AI Risk), developed by James Perkins at What About AI?, assesses AI career risk for 289 jobs across 26 industries. Each role receives two independent scores: a displacement score measuring the risk of being outcompeted by AI-literate peers, and a replacement score measuring the probability of full job elimination. Scores are derived from curated futurist predictions filtered for 90%+ historical accuracy, hands-on AI capability testing, industry adoption data, and peer-reviewed economic research.
Source: James Perkins, What About AI? ·
How we calculate AI displacement and replacement risk — and why the distinction matters more than any single “automation risk” number.
JATwo Scores, Not One
Most AI career predictions give you a single number. The FAIR Framework gives you two — because they tell fundamentally different stories.
Displacement Score
Measures how much harder your job becomes if you don't adapt. A high displacement score means AI-literate workers will increasingly outperform those without AI skills — even if the job itself isn't going away.
Think of it as competitive pressure. The job stays, but the requirements shift beneath you.
Replacement Score
Measures the probability that the role ceases to exist entirely. Full automation where AI or robotics can handle the complete scope of the job without human involvement.
Most jobs have high displacement but low replacement — the job stays, but the job description changes dramatically.
Why this matters: A Data Entry Clerk scores 98% displacement and 98% replacement — the role is being eliminated. A Surgeon scores 45% displacement but only 20% replacement — AI changes how the job is done, not whether it exists. Single-number frameworks would miss this distinction entirely. One should retrain. The other should upskill.
Data Sources & Inputs
Every score is grounded in multiple independent data streams. No single source dominates the output.
Futurist Evaluation
We systematically reviewed predictions from leading AI researchers, economists, and technology forecasters. Only those with a demonstrated accuracy rate above 90% on past predictions were included in our scoring inputs.
Industry Markers
Point-in-time industry adoption data — current AI tool penetration, automation pilots, regulatory signals, and workforce composition metrics — anchors each score to observable reality rather than speculation.
AI Capability Assessment
Hands-on evaluation of every major AI system's ability to perform the specific tasks within each role. Not theoretical benchmarks — actual capability testing against real-world job functions.
Economic & Labor Data
Research from Goldman Sachs, McKinsey, the World Economic Forum, the International Monetary Fund, PwC, and MIT provides the macroeconomic context that informs industry-level weighting.
Historical Pattern Analysis
Past technology transitions — from mechanization through digitization — reveal how job displacement actually unfolds. We weight for the speed differential: AI adoption is 5-10x faster than previous transitions.
Task-Level Decomposition
Rather than assessing jobs as monoliths, we decompose each role into its constituent tasks and evaluate AI applicability at the task level. This is what makes displacement and replacement scores diverge.
How We Evaluate Futurist Predictions
Not all predictions are created equal. We filter for track record before incorporating any forecast into our scoring.
Collect Predictions
We catalog specific, falsifiable predictions from AI researchers, economists, technology forecasters, and industry leaders — people like Geoffrey Hinton, Dario Amodei, Sam Altman, Elon Musk, Demis Hassabis, Ray Kurzweil, Andrew Ng, and Erik Brynjolfsson, among many others.
Verify Historical Accuracy
Each forecaster's past predictions are evaluated against outcomes. We specifically look for predictions with clear timelines and measurable criteria. Forecasters with less than 90% accuracy on verifiable past predictions are excluded from our input pool.
Cross-Reference with Ground Truth
Qualified predictions are cross-referenced against current industry adoption data, published AI benchmarks, and our own hands-on capability testing. A prediction that an AI system “can” do something is only weighted if we can verify the claim against actual performance.
Synthesize into Scores
Validated predictions, combined with industry markers, economic research, and our capability assessments, are synthesized into the displacement and replacement scores. The weighting formula is proprietary, but the inputs are documented here in full.
Five Risk Tiers
Each job's displacement score maps to one of five risk tiers, each with distinct strategic implications.
Roles facing near-term structural disruption. Significant task overlap with current AI capabilities. Immediate upskilling strongly recommended.
Substantial portions of the role are automatable within 2-5 years. Professionals should actively develop AI-complementary skills.
Meaningful AI impact expected. Some tasks will be augmented or automated, but human judgment remains central. Proactive adaptation advised.
Limited near-term disruption. AI will augment rather than replace. Roles benefit from AI literacy but face lower structural risk.
Roles with strong human-centric requirements — physical dexterity, deep empathy, unpredictable environments — that resist current AI capabilities.
Personal Risk Adjustment
Base scores reflect the role itself. Your personal score reflects you in that role.
Career Risk Quiz
A focused assessment that adjusts your base displacement score using three personal factors:
- •Years of experience — veterans with deep domain knowledge are harder to displace
- •AI exposure level — the single largest factor; active AI users face dramatically lower risk
- •Industry and role selection — maps to the base displacement score from our dataset
Transparency & Limitations
Honest assessment is a core value. That includes being honest about what our methodology can and cannot do.
What We Publish
- ✓All data source categories used in scoring
- ✓The complete five-tier risk classification system
- ✓Our futurist evaluation and filtering criteria
- ✓The displacement vs. replacement conceptual framework
- ✓All personal adjustment factors and their direction of impact
- ✓Full scores for all 289 jobs across 26 industries
Known Limitations
- •AI is evolving rapidly. Scores reflect current capabilities and near-term trajectories, not decades-out predictions.
- •Regional variation exists. AI adoption rates differ by country, state, and even city. Scores reflect aggregate trends, primarily weighted toward the U.S. and developed economies.
- •Regulation is unpredictable. Government action could accelerate or slow AI adoption in specific sectors in ways that are difficult to forecast.
- •Individual outcomes vary. A score is a starting point for planning, not a verdict. Your personal adaptability, network, and choices matter enormously.
- •Exact weights are proprietary. We disclose inputs and outputs transparently but do not publish the specific weighting formula, consistent with standard research practice.
Who Built This
What About AI? was founded by practitioners, not pundits.
James Perkins
Co-Founder · FAIR Framework Creator
25 years of experience across startups and corporate leadership, including financial operations at JP Morgan Chase. James developed the FAIR Framework and leads all data research, futurist evaluation, and scoring methodology behind What About AI?
Sean Boyce
Co-Founder
Background in technology with a focus on making complex topics accessible. Sean co-founded What About AI? and leads content strategy and practical AI implementation guidance, translating technical AI developments into actionable career advice.
Frequently Asked Questions
What is the FAIR Framework?
The FAIR Framework (Futurist-Informed AI Risk) is a methodology developed by James Perkins at What About AI? for assessing how artificial intelligence will impact specific careers. It produces two independent scores for each of 289 job roles across 26 industries: a displacement score (risk of being outcompeted by AI-literate peers) and a replacement score (probability of full job elimination by AI or robotics).
How does What About AI? calculate AI career risk scores?
The FAIR Framework evaluates both displacement risk and replacement risk using six data streams: curated futurist predictions filtered for 90%+ historical accuracy, hands-on AI capability testing, point-in-time industry adoption data, economic research from institutions like Goldman Sachs and McKinsey, historical technology transition patterns, and task-level decomposition of each role. The weighting formula is proprietary, but all inputs are documented publicly.
What is the difference between displacement and replacement?
Displacement measures competitive pressure — how much harder your job becomes if you don't adapt to AI tools. A high displacement score means AI-skilled workers will increasingly outperform those without AI literacy. Replacement measures full elimination — the probability that the entire role ceases to exist. Most jobs face high displacement but low replacement, meaning the job stays but the job description changes dramatically.
How many jobs and industries does the framework cover?
The current dataset covers 289 specific job roles across 26 industries, from Office & Administrative Support to Healthcare, Legal, Finance, Creative Arts, and more. Each job is individually scored rather than assigned a blanket industry average.
How often are the scores updated?
Scores are reviewed and updated as significant AI capability milestones occur or major industry shifts are observed. The underlying dataset was built through months of research and is maintained as a living document rather than a static snapshot.
Why don't you publish the exact formula weights?
We publish our data sources, scoring tiers, and the conceptual framework in full. The specific variable weights represent months of calibration work and are our proprietary methodology. This is consistent with how credit scoring agencies, economic forecasters, and research institutions operate — transparent inputs and outputs, proprietary weighting.
Can I personalize my risk score?
Yes. Our Career Risk Calculator lets you adjust your base score using six personal factors: years of experience, AI tool adoption level, education, work environment, management level, and whether your role is human-centric. Our Career Quiz provides a faster assessment using your industry, experience, and AI exposure level.
Who created the FAIR Framework?
The FAIR Framework was developed by James Perkins, co-founder of What About AI? James has 25 years of experience across startups and corporate leadership, including financial operations at JP Morgan Chase. He leads all data research, futurist evaluation, and scoring methodology. Sean Boyce co-founded What About AI? and leads content strategy and practical AI implementation guidance.