AI exposure
A measure of how much of a job consists of tasks that current AI systems can perform, used in labour market research.
Economists studying what AI does to work needed a way to sort jobs, and exposure is the measure they settled on. A job is highly exposed if a large share of its individual tasks are things a model can already do: drafting text, writing code, summarising documents, producing routine analysis. Software development, IT support, data work, marketing and finance tend to score high.
The trap is reading exposure as a forecast of job losses, which it is not. It says the tools overlap with the tasks, not what employers will do about it, and the evidence so far points both ways at once: highly exposed occupations have seen the fastest growth in advertised pay and a large share of layoffs that cite AI. A tool that overlaps with your work can replace you or make you considerably more valuable, and exposure alone cannot tell you which.