Will AI change your job—or just parts of it?

Start with the workday

Imagine an office worker who drafts messages, checks records and resolves awkward customer problems. That is an illustrative workday, not a measured occupation. A tool that helps with the first task has not demonstrated that it can take over the whole role.

What the research examined

The ILO’s May 2025 study assessed potential generative-AI exposure using task data, worker and expert input, and model predictions. It estimated that one in four workers worldwide was in an occupation with some exposure. Clerical occupations had the highest exposure levels. These are exposure estimates, not counts of jobs already lost. [1]

The opportunity—and the unresolved question

The authors expect transformation to be more common than complete replacement because many tasks still require human input. [1] Our interpretation: the practical question is how an employer reorganizes the work. Time saved on drafting could support more customer attention, or different staffing decisions. The exposure index alone cannot tell an individual which outcome will occur.

Look at tasks, not just titles

Try listing three things you actually do during a workday. For each, distinguish producing an output from checking it, handling exceptions and taking responsibility. This is a reflection exercise, not a personal job-risk score. It gives a more concrete starting point for a workplace discussion.

Original sources

Attributed synthesis, not original reporting. Examples labeled hypothetical or illustrative are explanatory. Reviewing a source does not independently validate its findings.

1. ILO Working Paper 140: Generative AI and Jobs ↗

   Published May 20, 2025. Publication summary and methodology description reviewed September 11, 2026; no independent replication.