How We Calculate AI Risk Scores

Our AI risk scores are derived from a systematic analysis of 40 work activities mapped across 6 dimensions of AI capability.

Data Sources

• O*NET 29.0 (2024) — 40 Detailed Work Activities
• Bureau of Labor Statistics OES — salary, employment, growth
• AI Capability Benchmarks (2025-2026) — MMLU, HumanEval, GPQA

The 6 Dimensions

1. Text & Language Processing (~20%)
2. Data Analysis & Pattern Recognition (~20%)
3. Visual & Creative Work (~15%)
4. Code & Logical Reasoning (~15%)
5. Physical & Manual Tasks (~15%, resilience)
6. Social & Emotional Intelligence (~15%, resilience)

Important Disclaimers

Task exposure does not equal full job replacement. Scores reflect current and near-term AI capabilities. Individual circumstances vary.

Last updated: March 2026. Next scheduled review: September 2026.