Artificial intelligence is the core of the digital economy. It is extensively applied in commercial and other settings, which has resulted in a series of challenges to international law. The governance need of artificial intelligence in technology and the governance effectiveness of sovereign states denote international law regulation as a necessary and feasible approach to artificial intelligence governance. At present, the international law regulation of artificial intelligence is generally immature and still in the primary stage, with its form and content respectively dominated by soft international law and indirect regulation. The challenge brought by artificial intelligence to international law in the context of digital economy is reflected both in the obsolescence of the current rules within international law and the limitations of international law regulation as a way of international governance. The reason for the former lies in the tension between exclusive sovereignty and unified digital space, which is manifested in the field of international economic and trade rules as the specific problems faced by artificial intelligence in its application layer, data layer and algorithm layer, including the unknown categorization of artificial intelligence products, the ambiguous legal nature of data, and the unclear regulatory approach of source code. The latter is due to the long period through which the rules of international law are formed, as well as the limited participants in the legislative process of international law. Correspondingly, the global cooperation of artificial intelligence governance centered around international law regulation should include both the updating of international law itself and its coordination with other international governance methods. The former indicates the evolution of traditional concepts in international law, including a series of evolution from Westphalian sovereignty to artificial intelligence sovereignty, as well as the coherence of the international legal system, which refers to the transition from existing rules to artificial intelligence rules based on the rules of digital economy. The latter not only requires the international legal governance of artificial intelligence to refer to and incorporate “super soft law”, but also requires the inclusion of multiple subjects in the international legal governance of artificial intelligence. As for China, at the domestic level, it needs to improve the artificial intelligence regulation system and enhance institutional stability; at the international level, it needs to participate in governance through international agreements, and encourage multiple stakeholders to participate in discussions over international governance of artificial intelligence at all levels.