Cognitive Automation and Sustainable Development: Global Task‐Level Evidence on Large Language Models and Labor Markets
Yuanfan Li, Rongrong Li, Qiang Wang · Sustainable Development · 2 Jun 2026
Where this sits in the debate
- Will AI destroy more jobs than it creates? · its finding: it depends (why studies differ)
Direction of the finding as read from the abstract by an AI model (AI-written).
Key takeaway AI-written
A bottom-up, task-level framework linking LLM capabilities to occupational structures across countries finds LLM-driven automation disproportionately exposes information-processing, administrative, and managerial roles, concentrating risk in knowledge-intensive sectors like finance and education while leaving agriculture and manufacturing comparatively insulated. It highlights a tension for sustainable-development policy: the sectors gaining the most economic value from AI are not the same as the sectors employing the most people.
Abstract
ABSTRACT The rise of large language models (LLMs) is transforming the trajectory of traditional rule‐based automation, while their global impact on the labor market remains largely unexplored. To assess this transition, we develop a novel bottom‐up framework linking detailed task data to occupational structures across a broad spectrum of economies. Our findings reveal that LLM‐driven automation disproportionately impacts roles centered on information processing, administration, and managerial coordination compared to those in physical or manual domains. At the sectoral level, this impact translates into higher exposure for knowledge‐intensive sectors like finance, education, and professional services, while sectors like agriculture and manufacturing remain more insulated. Because the industrial structure strongly determines a nation's vulnerability, economies reliant on administrative and clerical activities are facing greater exposure. Critically, potential productivity gains from this exposure are concentrated in areas contributing significantly to economic value rather than employment, highlighting a tension between efficiency and equitable labor outcomes. We argue that proactive policies, focusing on fair transitions, skill adaptation, and strategic industrial development, are crucial to avoid this technological shift undermining the sustainable development goals.
About this publication
Academic journal. Published in an academic journal — typically peer-reviewed before publication, though indexing alone doesn't guarantee review depth.
- Published
- 2026 (5 months ago)
- Citations
- 0 (OpenAlex)
0 citations recorded by OpenAlex — too few 2026 papers in this collection for a fair comparison.
Related by theme, not necessarily by method or conclusion.
- The Effects of Generative AI on High Skilled Work: Evidence from Three Field Experiments with Software Developers
SSRN · 2024 - GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models
arXiv · 2023 - Experimental evidence on the productivity effects of generative artificial intelligence
Science · 2023 - The simple macroeconomics of AI
Economic Policy · 2024
Overview assembled automatically from collected metadata. How this site is built.