Check the AI Automation Risk for Your Profession
Free tool that estimates AI automation risk for any occupation. It blends real Claude usage patterns from the Anthropic Economic Index, BLS Employment Projections 2024-2034, PayScope 2026 exposure index, and O*NET task statements into a composite score with an honest confidence range. Type your job title, see the risk gauge, the time-horizon curve, and a task-level breakdown showing which parts of the work AI can already do.
How to read your risk score
Use the search box or pick from the trending tile. Autocomplete narrows 50+ professions as you type.
The center number is the 5-year exposure. The thin arc shows the disagreement between data sources — wider is less certain.
Compare 1-year, 5-year and 10-year risk. The 10-year value extrapolates the Anthropic adoption curve and carries more uncertainty.
Each task is colored: red = automatable today, amber = augmented by AI tools, green = still human-led. Plan reskilling around the green and amber pools.
The "Related" panel shows neighbors in the same SOC family with their risk scores — useful for considering a pivot.
See how exposed your profession is, with a confidence range from real 2026 datasets.
Suggest a profession We blend US BLS occupational codes — modern niche roles (SEO, DevOps, SMM…) may not be in the taxonomy yet. Tell us and we'll map them to the closest BLS family.
🔥 Trending searches
⚠️ Highest automation risk
🛡️ Lowest automation risk
No measured Claude usage
The Anthropic Economic Index has not recorded Claude usage for this occupation. This typically means: hands-on physical work, in-person interaction, or fewer text-driven tasks. The 0% score should be read as "no AI engagement measured today", not "AI cannot help".
Pick from related or type above.
Share of measured tasks where AI assists the worker (feedback, learning) rather than replaces them (directive, validation). High augmentation + low risk = AI changes the job without threatening it.
Safer roles with related skills When the composite risk is high, these are lower-exposure peers in the same industry — ranked by skill-transfer proxy (closer SOC stem first).
Skills AI can't easily replicate Where to invest if you want to stay ahead in this profession. Combination of your own task breakdown plus traits that stay human-led across the industry.
Task-level breakdown How AI splits across the typical duties of this role today.
Related occupations Compare risk across the SOC family.
How predictions aged Frey & Osborne (2013) vs. modern multi-source 2026 estimate. Big gap = the 2013 model missed the LLM wave.
No 2013 prediction for this occupation — Frey & Osborne sampled 702 SOCs, this one was not in the original set.
Sources
Composite risk blends three primary datasets, joined by SOC code. We do not invent the numbers; each source links out below so anyone can audit.
Anthropic Economic Index (release 2026-03-24, "Learning curves"): real Claude usage patterns by O*NET-SOC 2019 occupation, including a 5-way interaction split that we collapse into the composite as the primary risk driver. view source ↗
BLS Employment Projections 2024–2034 (released Aug 28, 2025): wages, growth, openings, and the 832-SOC base we anchor everything to. view source ↗
BLS SOC 2010→2018 crosswalk: used to align AEI's task statements (still in O*NET-SOC 2010) with the 2018 BLS SOC employment file. view source ↗
O*NET 30.3 task taxonomy (May 2026): the underlying task descriptions ride along inside the AEI download. view source ↗
Frey & Osborne (2013), "The Future of Employment": kept only as the historical baseline for the perception-drift chart, NEVER in the composite. view source ↗
Estimates are based on US labor data. Real exposure depends on your specific employer, region, and tasks. Treat the score as a planning aid, not a forecast.
