At a glance
- Source
- 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 doi:10.54394/fhem8239 · Not peer-reviewed: other researchers have not yet checked it for a journal.
- How it was done
- The authors scored the tasks in each occupation for how far a generative AI model (GPT-4) could do them, to estimate how exposed occupations are around the world. Data analysis. It builds a measure from existing data, such as job and task records.
- Who took part
- Occupations and their tasks, across countries at all income levels.
- Where
- Worldwide.
“It predicts that the overwhelming effect of the technology will be to augment occupations, rather than to automate them.”
What it found
- The main effect of generative AI is likely to be assisting people in their jobs, not automating the jobs away.
- Clerical work is the most exposed.
- The effect is largest in high-income and upper-middle-income countries, where more people do clerical work.
- Clerical jobs employ many women, so the effect differs sharply by gender.
When the result holds
The estimate describes what the technology could do. Whether a job is assisted or replaced depends on how employers choose to use it.
Limits
- A forecast built on task scores, not a record of what happened.
- The scores reflect one AI model in 2023.
Practices that use this study
This study supports: