aceModerate evidence
Published recordLaborProblem Space
More
Does artificial intelligence impact employment in the medium-skilled labor force? Evidence from China’s National AI Innovation and Application Pilot Zones
- Truvace record
- TRV-2026-2517raw data (JSON)
- Version
- v1Published
- First published
- Evidence type
- Observational study
- Sources
- 1 sourcePeer-reviewed
- sha256
- 58cbdaf96129…check it
Rapid artificial intelligence diffusion in China, shaped by government initiatives, is reshaping labor markets and raising questions about medium-skilled employment. Using a 2015–2023 panel of 287 Chinese prefecture-level cities, we examine how policy-induced artificial intelligence diffusion affects medium-skilled employment. Identification exploits the staggered establishment of China’s National AI Innovation and Application Pilot Zones as a quasi-natural experiment in a staggered difference-in-differences des…
The problem
A 2015–2023 study of 287 Chinese cities found that cities designated as national AI pilot zones saw a negative, statistically significant effect on medium-skilled employment compared with other cities.
What this doesn’t fix
The study measures city-level employment effects of a policy designation, and the authors note that effects differ significantly across treatment cohorts and are non-monotonic over time.
Sources
1 on file- Journal of the Asia Pacific Economydoi:10.1080/13547860.2026.2741537
Cite this record
v1Truvace Impact Record TRV-2026-2517, v1: “Does artificial intelligence impact employment in the medium-skilled labor force? Evidence from China’s National AI Innovation and Application Pilot Zones.” Truvace, 2026-11-12. /record/TRV-2026-2517 (accessed on the day you copy it). sha256 58cbdaf961299389…
Change history
1 versionEvery change made to this record since it was first published, and why. Nothing has changed since it went up.
Certified into the record
Check this record hasn’t been altered
How to check it yourself, without trusting this page
Download any version from /api/record/TRV-2026-2517 and hash its canonical field yourself, for example with shasum -a 256. You should get its content_hash. Each version’s text also ends with prev:and the hash of the version before it (version 1 points to 64 zeros). If anyone had changed a single character of any version after it was saved, those hashes wouldn’t line up.