Not yet human-reviewed
This source record was written automatically by the research pipeline and has not yet been reviewed by the maintainer or integrated into a topic page. Treat it as a provisional summary and verify claims against the original source.
AI and Workforce Dynamics: Unravelling Productivity
Citation. Yoshida, H. & İnce Yenilmez, M. (2025). AI and Workforce Dynamics: Unravelling Productivity. Sosyoekonomi, 33(63), 143–159. https://doi.org/10.17233/sosyoekonomi.2025.01.07
Summary. This peer-reviewed article models the association between firms’ AI-related patents and labour productivity, alongside non-AI patents, R&D investment, physical capital, and employment. Its reported estimates associate more AI-related patents with higher productivity, while the design is observational and cannot establish that patents themselves caused the productivity change.
Key claims
- In the reported random-effects model, AI-related patents were positively associated with labour productivity (coefficient 0.318, p < .01), while non-AI patents also had a smaller positive association (0.307, p < .01).
- The analysis reports 363 observations and finds inverse associations between employment count and labour productivity, which the authors interpret alongside possible production and displacement effects.
Evidence & limitations
- The paper analyzes a balanced firm panel described as covering 2020–2023 using patent counts and productivity-related covariates; its observational design does not isolate causal effects of AI adoption.
- The sample includes firms with at least one AI-related patent, so selection and the use of patents as an AI measure limit generalization to firms without AI patents or to actual workplace use.