AI job exposure statistics: how exposed people say their weekly tasks are
This page publishes once at least 30 people have completed the AI Task Exposure Check in the current window. Until then the count is shown as below the reporting floor and no figures are published.
Once a cut reaches 30 completions, it shows how completions split across the four levels (Task Safe, Task Shifted, Task Exposed, Role Exposed), the mean on each dimension, the share whose Replaceable Tasks score is above the midpoint, which is the closest this check comes to saying more than half the week is replaceable, and the same figures for respondents who manage people. Every figure describes how respondents see their own tasks. None predicts a job.
Leaders who score their own AI readiness are reported separately on the AI leadership benchmark. The manager cut here can be read next to it, in words, and is never merged with it: the two pages ask different people different questions.
Current edition
Fewer than 30 completions in this cut. Nothing is published below that floor. Next refresh: 2026-10-12.
Collection started on 2026-09-22. This page is rewritten when a refresh runs; the change log below carries every edition.
What this page will report
So that no cut is chosen after seeing the data, here is what is published when n reaches 30, and nothing else:
- Level distribution (share of completions in each band)
- Replaceable over half share
- Mean score per dimension, 0 to 100
- Level by role level
- Region
The floor rule
No cut is published with fewer than 30 completions. Where a cut falls below the floor, the chapter is published with the count and no figures. The floor is set before the edition and never lowered. Completions are counted, not people: the capture is anonymous and cookieless, so a returning visitor is not identifiable and is not de-duplicated. Ad blockers drop some events, so every count is a floor. The full rules are on the methodology page.
What this page will never publish
- Any prediction that a job, an occupation or an employer will be replaced by AI. The unit is a respondent's own reading of their week.
- Any figure merged with the AI leadership benchmark. The two populations differ and are compared in words only.
- Any cut that names or could identify an employer, or an industry small enough to identify one.
- Any cut below 30 completions, or any cut that still contains the site's own test runs.
How to read it against other surveys
The outside figure to read this against is Pew Research Center's survey of 5,273 employed US adults in October 2024: 52 percent said they feel worried about how AI will be used in the workplace in the future, and 16 percent said they already use AI for some of their work (Pew Research Center). Pew reports how people feel and whether they use AI. This page reports which of their own tasks people see as exposed, and splits it by whether they manage people, which Pew does not publish.
How to cite this page
Leaders ADAPT, AI job exposure statistics: how exposed people say their weekly tasks are, Workplace Reality Index, edition 2026-09-23, https://www.leadersadapt.com/ai-job-exposure-statistics/. Until a cut passes the floor there is no figure to cite; cite the method instead. Every published figure will carry its n, its window and its date on the same line.
Change log
| Date | Change |
|---|---|
| 2026-09-22 | Page created with the tool. Collection begins. No figures published; n below floor. |
Take the tool
Every completion of the AI Task Exposure Check: how much of your job is AI taking joins this sample anonymously. More free tools are at the tools shelf.
The four results
Every completed run lands on one of these pages, with your own scores drawn from the link the tool gives you. Each one is readable without taking the tool.
- Task SafeLittle of your week is exposed. Your judgement carries the job.
- Task ShiftedPart of your week is shifting to tools. The core stays yours.
- Task ExposedMost of your week is exposed. Time to reshape it yourself.
- Role ExposedMost of your week is exposed and decided elsewhere. Start with Direction.
The other statistics pages
Every open tool on this site publishes its own statistics page under the same floor rule:
- Am I the problem at work? Statistics on self view and how others see you
- Boss pattern statistics: what people report about their bosses
- Golden handcuffs statistics: what people say keeps them in their jobs
- Job security statistics: the signals people report around their jobs
- Micromanaging statistics: how closely people say they are managed, and whether they know why
- Promotion block statistics: what stands between people and the next role
- Become a manager statistics: who wants the work, and who wants the title
- Should I quit statistics: what people asking are really leaving
- Toxic workplace statistics: four conditions at work, with the manager view first