Overview
This series explains why AI exposure is a capability-distance signal, not a forecast of job loss. It moves from the OECD’s nine-domain gap index and judgement-based occupational map, through alternative frontiers and a one-level cognitive scenario, to 2024 European survey evidence from about 36,600 workers in 35 countries. Exposure is associated with reported use, but skills, digital access, training, workplace influence, institutions, and country context shape the translation. The study found no detectable average early-adoption effect on worker-reported task displacement or creation, so redesign remains a separate, time-dependent question.
Learning outcomes
- Exposure measures proximity between AI capabilities and occupational requirements; it does not measure adoption or job loss.
- The OECD’s broad nine-domain profile reveals different bottlenecks, so one exposure ranking cannot represent every work pathway.
- Exposure is strongly associated with reported AI use, but skills, digital readiness, training, and workplace influence shape adoption.
- A demanding European test found no detectable average early-adoption effect on worker-reported task displacement or creation.
Keywords
Preview every chapter · Unlock full text, audio, and video from chapter 4
PlansFAQ
What does an AI exposure score actually measure?
It measures how close assessed AI capabilities are to the capabilities an occupation requires, not whether the job will be replaced. The OECD index leaves adoption, cost, and institutional feasibility unresolved.
Why can two AI exposure rankings disagree?
They can disagree because they compare different AI capability frontiers with different parts of an occupation’s profile. A language-focused measure may highlight document tasks, while the OECD measure also counts social, sensory, physical, and robotic capabilities.
What determines whether occupational exposure becomes workplace AI adoption?
Exposure becomes workplace adoption only when workers and organisations can act on it. The 2024 European study links uptake with education, abstract task content, organisational influence, digital readiness, and training conditions, but does not establish each relationship as causal.
Does AI adoption prove that a job’s tasks have changed?
No; AI adoption does not by itself prove that a job’s task bundle has changed. Adopters reported more task displacement and creation descriptively, but the more demanding estimates found no detectable average early-adoption effect.