Open-source organizational modes, innovation costs, and perceived policy effects

Open-source innovation, as a pivotal organizational model in the artificial intelligence (AI) ecosystem, holds strategic significance for technological advancement and high-quality innovation. With the global adoption of open-source large AI models such as DeepSeek and Qwen, China’s open-source organizational models have drawn increasing international attention. Yet, scholarly understanding of how these organizations perceive the effectiveness of national open-source innovation policies-and the underlying mechanisms-remains underexplored. Drawing on a nationwide survey of Chinese open-source organizations, this study examines the relationship between organizational modes of open-source innovation and policy effectiveness perceptions. Findings reveal that, under broadly implemented national policies, organizations with lower levels of hierarchical structure report significantly stronger policy effectiveness perceptions. This relationship is mediated by innovation-related costs: behavioral incentive costs and output appropriation costs exhibit positive mediating effects, indicating that moderate governance investment and clear attribution of outputs enhance policy perception; in contrast, regulatory coordination costs show a negative mediating effect, reflecting how institutional friction and coordination burdens diminish perceived policy efficacy. Notably, output appropriation costs exert a particularly pronounced influence on perceptions of funding allocation. These results uncover the heterogeneous roles of distinct innovation costs in shaping policy perceptions, highlighting opposing directional effects of governance-oriented versus coordination-oriented costs. Empirically grounded in China’s open-source practices, the study further demonstrates that open-source governance is not merely community self-governance but rather an institutionally embedded hybrid governance model. The findings offer valuable theoretical and practical insights for refining policy design, optimizing organizational governance, and fostering the development of AI innovation ecosystems.