An estimate, not a prediction. This is an estimate of how much of this kind of work AI tools can do or speed up, built from published task lists. It is not a prediction about your job, your employer or your career: most jobs change rather than disappear, and the parts that change are often the routine ones.
How we work this out
Every task O*NET lists for the occupation is put in one of four groups by its main verb and what it acts on. Information work counts fully towards exposure, planning and deciding half, working directly with people a fifth, and hands-on physical work not at all. Tasks are weighted by how important O*NET's survey of workers rates them. The weighted result is shown only as a band.
- Information work (reading, writing, calculating, analysing)
counts fully towards exposure - Planning, deciding and coordinating
counts half towards exposure - Working directly with people
counts a fifth towards exposure - Hands-on physical work
counts not at all towards exposure
Bands: Lower exposure, Medium exposure from 40 in 100 weighted task points, Higher exposure from 65 in 100 weighted task points. Supplemental tasks count 70% of a core task.
Demand: new postings JobRipe first saw from 2026-07-27 to 2026-10-25 compared with the 90 days before (2026-04-28 onwards), only from employers JobRipe was already tracking before 2026-04-28, so adding employers never looks like growth. A trend needs at least 20 earlier postings from 5+ employers; a change of 15% or more is “growing” or “shrinking”. A rising or falling skill needs 10+ postings and a 5-point move.
Sources
- O*NET 30.3 Database (task statements, task importance ratings, occupation data), U.S. Department of Labor, Employment and Training Administration (USDOL/ETA) · CC BY 4.0
- Autor, Levy & Murnane (2003), The Skill Content of Recent Technological Change, Quarterly Journal of Economics 118(4) · used for the routine / non-routine task framework (method only, no data)
- Eloundou, Manning, Mishkin & Rock (2023), GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models · used for the idea of rating exposure task by task from O*NET (method only, no data)
- Gmyrek, Berg, Kamiński, Konopczyński, Ładna, Nafradi, Rosłaniec & Troszyński (2025), Generative AI and Jobs: A Refined Global Index of Occupational Exposure, ILO Working Paper 140 · used for exposure shown as gradients rather than a single number; transformation, not replacement, as the likely effect (method only, no data)
- JobRipe job postings · used for demand trend, rising and falling skills, open jobs (same-panel employers only)
This page includes information from the O*NET 30.3 Database by the U.S. Department of Labor, Employment and Training Administration (USDOL/ETA). Used under the CC BY 4.0 license. O*NET® is a trademark of USDOL/ETA. JobRipe has modified all or some of this information (task grouping and exposure bands). USDOL/ETA has not approved, endorsed, or tested these modifications.