Researchers from University of Pittsburgh, University of Milan, and Berlin School of Economics analyze German Socio-Economic Panel data to assess AI exposure’s impact on worker wellbeing and health. Using event study and difference-in-differences methods, they compare high- and low-AI-exposure occupations before and after 2010. Findings show no negative effects on life or job satisfaction, and modest improvements in self-rated health and health satisfaction, possibly due to reduced physical strain.
Key points
- Combines the Webb (2019) occupational AI exposure index and a SOEP-based self-report metric to classify AI exposure levels.
- Implements event study and DiD models with individual, state-year, occupation, and industry-year fixed effects to isolate AI’s causal impact.
- Finds no significant negative effects on life satisfaction, job satisfaction, mental health; reports modest self-rated health and health satisfaction improvements.
Why it matters: Revealing AI’s neutral effect on wellbeing and modest health gains provides evidence for workplace AI policies that protect employee health.
Q&A
- What is the Webb AI exposure measure?
- How do event study and difference-in-differences methods work?
- Why use self-reported health and satisfaction metrics?
- How can AI adoption lead to improved worker health?