Free tool. No signup. Runs in your browser.

Will AI Take My Job? Task Exposure Checker

Pick your occupation or list your weekly tasks and see which parts AI can already do, which it cannot, and what to learn next.

67 occupations, each with five typical weekly tasks.

Ratings are editorial estimates of what general-purpose AI assistants do well in 2026, task by task. They describe work, not people, and are not designed for or suitable for decisions about any individual's employment.

Your task-by-task breakdown appears here

An exposure score, each task rated as AI-producible, AI-assisted or person-only, and the three skills that move you toward the person-only side.

How to use the AI Job Checker

  1. 1

    Pick your occupation, or list your tasks

    Search the library or describe a real week in your own words, one task per line. No names or personal details, only tasks.
  2. 2

    Read the task breakdown

    Each task is rated AI-producible, AI-assisted or person-only, with a one-line reason. The score summarises the mix; the rows explain it.
  3. 3

    Move toward the person-only side

    The three skills at the bottom are chosen to make you the person who runs the AI-producible tasks and owns the person-only ones.

Next step

Be the person who runs the AI, not the one measured against it

Check your occupation or task list above and this block turns the result into the two skills to learn first. Coursiv teaches them in role-based courses.

See my learning plan

A short quiz first. No payment details.

Will AI take my job? The task view is the honest answer

Jobs are bundles of tasks, and AI adopts tasks one at a time. A data entry clerk and an electrician can both list "writing reports" among their tasks; for one it is most of the week, for the other it is an hour on Friday. Asking whether AI will take a job hides that difference. Asking which tasks it can already produce makes the answer specific, and specific answers are the ones you can act on.

The pattern across the library is consistent: writing, summarising, scheduling, data handling and first drafts of almost anything rate high; analysis, planning and troubleshooting rate medium because a person still decides; and anything that needs presence, hands, trust or accountability rates low. Most jobs contain all three.

What a high exposure score means in practice

In roles where most tasks are AI-producible, the people who learn to brief, check and combine AI output early tend to absorb work that used to take several colleagues, and the role gets measured against that pace. The risk is not the tool; it is being the person in the team who does not use it. A high score is therefore a reason to learn the tools this quarter, not next year.

What this tool is not

It is not a prediction about any employer, a measure of any individual, or an input to decisions about people. It reads task descriptions only, refuses anything that looks like personal data, and rates categories of work with editorial estimates that will age as models improve. Use it the way you would use a map of a moving landscape: to decide which direction to walk.

Preguntas frecuentes

Each task is rated high (current AI assistants can produce it from a short brief), medium (AI drafts or suggests, a person decides) or low (needs a person present, hands-on or accountable). High counts as 1, medium as 0.5 and low as 0.1; the score is the average across the tasks, as a percentage.