Conflicts of Jurisdiction in Artificial Intelligence-Related Intellectual Property Disputes: Towards Coherent Private International Law

The cross-border development and commercialisation of generative artificial intelligence (AI) raise jurisdictional questions with implications for intellectual property protection and innovation governance. Because AI activities span data collection, model training, and subsequent exploitation, disputes may involve distinct claims connected to multiple jurisdictions. This article examines these questions through a comparative doctrinal analysis of South Korea, Japan, China, and Vietnam. It compares the four systems using six jurisdictional criteria: defendants, local business activities, contracts, infringement, registered rights, and the coordination of concurrent proceedings. The analysis shows that the jurisdictional significance of territorial connections depends on the particular claim and the legal ground invoked. The location of technical infrastructure does not, by itself, establish jurisdiction over every dispute concerning a model. Even where local exploitation provides a basis for jurisdiction, that jurisdiction cannot automatically extend to earlier training abroad. Nor do proceedings involving the same system necessarily constitute duplicative litigation, since they may concern distinct rights and acts. The article argues that the AI lifecycle should guide the identification of particular claims and the coordination of overlapping proceedings, rather than constitute an autonomous ground of jurisdiction. By distinguishing jurisdiction from the territorial requirements of infringement, the proposed approach clarifies the limits on concentrating claims in a single forum. It offers a framework for applying existing rules more coherently, with the aim of improving legal predictability for rights holders, AI developers, and businesses engaged in cross-border innovation. Its practical effectiveness requires further examination of judicial practice.