Research on Data Intellectual Property Protection and Value Assessment in the AI Era: A Tri-Jurisdictional Comparative Study of China, the United States, and the European Union

Abstract

As artificial intelligence reshapes the global economic infrastructure, data has ascended from a by-product of digital activity to the primary factor of production in the twenty-first century. Yet the very legal regimes tasked with protecting data-derived value remain profoundly fragmented. This paper conducts a comparative institutional analysis of how China, the United States, and the European Union construct data intellectual property (IP) protection frameworks, and how those divergent architectures interact with market mechanisms for valuing data assets. Drawing on three paradigmatic cases—(1) hiQ Labs v. LinkedIn (U.S.), (2) the EU's sui generis database right regime under the Database Directive and subsequent Data Act reform, and (3) China's emerging data-IP registration pilot system exemplified by the Shanghai and Zhejiang platform-data disputes—this study identifies systematic patterns: the U.S. relies on a decentralized, litigation-driven mosaic (trade secrets, CFAA, contract, copyright) that maximizes allocative flexibility at the cost of transactional uncertainty; the EU deploys a rights-centric, regulatory-dense architecture (database right + GDPR + Data Act) that secures individual and producer protections but imposes heavy compliance frictions on data monetization; and China pursues a state-guided, registration-enabled property-rights experiment ("three-rights" bifurcation + data IP pilots) designed to engineer a tradable data-asset class from above. The paper then develops an interaction-mechanism framework showing how each jurisdiction's legal DNA shapes which valuation methodologies (cost / income / market) dominate, what discount rates the market applies, and where liquidity emerges or stalls. Business and policy implications are drawn for firms operating across jurisdictions in the AI supply chain.