About This Project
The AI Work Index scores 562 occupations and 80 modern roles for AI displacement risk. It uses a three-layer model — exposure, human bottleneck, and market resilience — grounded in peer-reviewed academic indices and official Singapore government data.
This is a structural score of AI pressure, not a prediction of exact job losses. No LLM is used in the scoring pipeline. The model separates technical exposure from labour-market displacement — a software developer and a data entry clerk can both score high on AI exposure yet have very different outcomes.
This model measures one side of the equation
In the Acemoglu & Restrepo (2019) framework, AI's net impact = displacement − reinstatement. We measure displacement only. Scores likely overstate net risk for occupations where AI creates new work.
State of the science (early 2026)
- Single exposure scores are poor unemployment predictors — ensembles do better (Frank et al., 2025)
- No consensus on measurement — "still in the first inning" (Brookings/PIIE, 2026)
- Entry-level workers face earliest pressure (Stanford DEL, 2025; Anthropic, 2026)
Model Card
Exact / Validated
- AIOE exposure scores (peer-reviewed, deterministic)
- Theta complementarity scores (O*NET survey data)
- Net risk formula (fully reproducible)
- Official demand signals (SOL 2026, Jobs in Demand)
Estimated / Group-Level
- Market resilience (group-level employment + wage heuristics)
- Crosswalk quality (US occupations mapped to SG)
- Labour monitor (cluster-level, not occupation-level)
- Anthropic calibration (Claude usage, not universal AI)
Synthetic / Illustrative
- Modern role estimates (weighted SSOC blends, medium confidence)
- Transition scores (heuristic feasibility estimates)
- Outlook/scenario modelling (rule-based, not predictive)
- Seniority modifiers (research-grounded, not validated)
Not Validated
- Occupation-level backtesting (cluster-level only: 3/4 pass)
- Company-size modifiers (not yet implemented)
- Causal displacement claims (directional correlation only)
- Occupation-level employment data (not publicly available)
Data Vintage
Wages
2024 MOM data
Demand Signals
SOL 2026 + Jobs in Demand 2025
Labour Market
Q3 2025 full + Q4 2025 advance
Model Version
V3.1 · 48 checks
Inspiration & How We Differ
Inspired by Andrej Karpathy's AI Job Exposure Map (March 2026) and Josh Kale's extended visualization, which score 342 US occupations using LLM-generated ratings (Gemini Flash, 0–10 scale).
What we do differently:
- No LLM in scoring — we use peer-reviewed academic indices (Felten AIOE, Pizzinelli theta), not LLM-generated subjective ratings
- Singapore-specific — SSOC occupational classification, MOM demand signals (SOL 2026, Jobs in Demand), Singapore labour market data
- Three-layer decomposition — we separate AI exposure from human bottleneck from market resilience, not a single composite score
- Validated — cluster-level backtesting against actual labour outcomes (3/4 directional checks pass), 48 structural checks
- Seniority modifiers — research-grounded experience level adjustments (Stanford DEL, Anthropic 2026)
- 80 synthetic roles — modern job titles (AI Engineer, Prompt Engineer) scored as weighted SSOC blends
License & Credits
MIT License. Adaptable for other countries via ISCO-08 crosswalks.
Made by Kirill So with Claude (Anthropic) & Codex (OpenAI). Data from MOM, Felten et al. (2021), Pizzinelli et al. (2023), O*NET, Anthropic Economic Index, and Stanford DEL.