AI-first company

Assessment · Task and job maps

ILO index of occupational exposure to generative AI

In which roles could generative AI take over tasks, and in which could it assist?

By
Last reviewed
Corrections
Report an error
Who answers Nobody is surveyed
How long You match your job titles to a standard list of occupations
Can you use it? Free to use and adapt
Testing Emerging

Early testing only: a pilot, an expert review, or tests that found problems still to be fixed.

In short

A public score of how far generative AI could do the tasks in each occupation, and whether it would more likely replace or assist the worker. Use it to choose where to look first. It is not a decision.

For whoever runs the assessment: an HR or operations lead, an internal project team, or an outside consultant.

What it measures

For each occupation in the international standard list (ISCO-08), the index looks at the tasks that make up the job and scores how far generative AI could do each one. From those scores it separates two cases:

  • Automation potential: most tasks in the job could be done by AI.
  • Augmentation potential: some tasks could, and the rest still need the person.

The study behind it is described on its own page. Its main result is that assisting people is the common case and replacing them the rare one.

How to use it

  1. Match your job titles to the occupation list.
  2. Use the scores to pick two or three roles to look at closely.
  3. Then map the real tasks in those roles with the people who do them. See map tasks before deciding what to automate.
  4. Decide who sees job-level scores and how they are explained. Telling people their job is “exposed” raises fear, and fear has been linked to staff hiding what they know.

What it does not measure

  • Use. Exposure is what AI could do. What AI is used for is a different thing, and predicts demand for workers better.
  • Your jobs. The scores describe a typical occupation. Your planner or accountant may do other tasks.
  • Knowledge that is hard to write down. Experience, judgement and relationships are not in the task lists.

How it was tested

The scores in the 2023 study were produced by an AI model (GPT-4) rating each task. A later national version, built with Poland’s research institute NASK, goes further:

  • it covers more than 2,500 occupations and close to 30,000 tasks in the national job classification;
  • it adds ratings of tasks by people and a survey of workers;
  • it gives each task a score from 0 to 1, and publishes the full index.

That makes the method transparent and better grounded. It is still a judgement of what the technology could do.

Licence and access

The ILO report is open under CC BY 4.0. The national version for Poland is published under CC BY-NC 4.0: free to use and adapt with credit, but commercial use, such as selling an assessment built on it, needs the publishers’ consent.

Alternatives

Other assessments with a similar job, and when to prefer each.

Practices that use it

The details

Testing
Emerging The method is published and has been extended in a national version that adds human ratings and a survey of workers. No study shows that the scores predict what happens to jobs or tasks inside a company.
Looks at
Job and its tasks
How the answers are collected
Scores given to the tasks of each occupation. No staff time is needed.
Length
No survey. You match your job titles to a standard list of occupations.
Cost
Free
Licence
CC BY 4.0You may use and adapt the index if you credit the source. A national version for Poland has a non-commercial licence: see below.
Languages
English. A national version exists in Polish.
By
Gmyrek, Berg and Bescond, International Labour Organization, 2023

Sources

  1. Gmyrek, Berg and Bescond (2023). Generative AI and jobs: A global analysis of potential effects on job quantity and quality. International Labour Organization. doi:10.54394/fhem8239
  2. Troszyński, Berg, Gmyrek, Kamiński, Konopczyński and colleagues (2025). Generatywna sztuczna inteligencja a polski rynek pracy. Raport badawczy. NASK – National Research Institute. nask.pl

Related studies on this site