How China Governs Open Science: Policies, Priorities, and Structural Imbalances[v1] | Preprints.org

Abstract:  This article investigates the architecture and institutional distribution of policy tools supporting Open Science (OS) in China. Based on a corpus of 199 policy documents comprising 25,885 policy statements, we apply an AI-assisted classification approach to analyze how the Chinese government mobilizes different types of policy tools. Using Qwen-plus, a large language model developed by Alibaba Cloud and fine-tuned for OS-related content, we categorize each policy statement into one of fifteen subcategories grouped under three main types: supply-oriented, environment-oriented, and demand-oriented tools. Our findings reveal a strong dominance of supply-oriented tools (63%), particularly investments in infrastructure, education, and public services. In contrast, demand-oriented tools remain marginal (11%), with limited use of economic incentives or mandatory mechanisms. Environment-oriented tools are more balanced but still underrepresent important elements such as incentive systems and legal obligations for open access. To deepen the analysis, we propose a normalized indicator of institutional focus, capturing the relative emphasis of each policy type across administrative levels. Results show that supply-oriented policies are most concentrated at Level 1 institutions, reflecting the Chinese government’s top-down approach to science governance. Environmental-oriented tools, by contrast, are more evenly distributed, while demand-oriented tools are predominantly localized at Level 4, suggesting limited strategic planning at higher levels. Taken together, the results highlight a technocratic and centralized model of OS governance in China, prioritizing infrastructural and procedural reform over participatory or incentive-based approaches. The unbalanced policy mix and weak demand stimulation may hinder the systemic transformation required for a sustainable and inclusive open science ecosystem.