数据知识产权保护与企业数据技术创新——来自生成式大语言模型的经验证据

Enterprise data technology innovation is a crucial process for realizing the value conversion of data elements. However, there is still a lack of systematic empirical research on how to effectively incentivize it through data intellectual property protection. Based on data from China’s A-share listed companies from 2015 to 2024, this paper innovatively employs a generative large language model to identify enterprise data technology patents. Using the first batch of national pilot projects for data intellectual property as a quasi-natural experiment, it constructs a DID model to systematically assess the impact of data intellectual property protection on enterprise data technology innovation. The results show that the data intellectual property pilot significantly promotes enterprise data technology innovation. Mechanism testing indicates that this effect is primarily realized through incentivizing specialized factor inputs and facilitating knowledge spillovers. Heterogeneity analysis reveals that the promoting effect is more pronounced in state-owned enterprises, data-native enterprises, and regions with better-developed digital infrastructure, while no significant difference is observed among enterprises with different technological innovation attributes. Further research demonstrates that by fostering enterprise data technology innovation, data intellectual property protection effectively enhances enterprise TFP. This paper provides new evidence for understanding the micro-level innovation effect of data intellectual property protection and offers important policy implications for building a unified national data market, cultivating new quality productive forces, and promoting high-quality development of the digital economy.