AI may change your job before it takes your job.

A job is a bundle of tasks, responsibilities, relationships and experience. AI can automate some tasks and increase the value of others. The early signs of change may therefore be altered hiring, slower career progression, different demands on junior staff or constant pressure to produce more — long before an occupation vanishes.

The IMF’s widely cited estimate that almost 40% of global employment is exposed to AI describes the reach of a technology, not how many jobs will be lost. Some workers may become more productive and better paid; others may face lower demand or reduced bargaining power. The direction depends on adoption, economics and policy.

Junior roles often contain routine work that helps new workers build judgment and credibility. When that work is automated, employers may benefit immediately, but the future pipeline of skilled professionals can weaken. A good labour-market dashboard should therefore follow entry-level opportunities and mentorship, not just top-line unemployment.

Where are AI tools changing the tasks within occupations? Are wages rising with output? Which regions and workers lack bargaining power? Are productivity benefits financing better public services or flowing primarily to owners of scarce assets? Good policy begins with measurement that can disprove a preferred narrative.

AI could produce a genuine public dividend: fewer dangerous tasks, more time for care, accessible services, and work with more meaning. Achieving that future requires stronger transition support, worker voice, competition and explicit decisions about distribution. A faster economy does not, by itself, make a fairer society.