EU Artificial Intelligence Act Article 50 Rules and Their Implications for the Construction of China’s Artificial Intelligence Legal System

Abstract

In August 2026, the European Union published its Artificial Intelligence Act and formally commenced its implementation, marking the first comprehensive legislation on AI oversight and control to enter the enforcement phase. The Act centres on “mandatory disclosure of AI identity, deep labelling of generated content, and risk-based systemic prevention and control,” and establishes a full-chain, all-encompassing, operable, and verifiable rule system. Meanwhile, China is at a critical juncture of moving toward comprehensive AI legislation. Notably, the 2026 Legislative Plan of the State Council made arrangements to accelerate comprehensive legislation for the sound development of AI. This paper systematically reviews the institutional structure and implementation logic of the transparency rules under the EU Artificial Intelligence Act, then analyses the existing framework and legislative progress of China’s AI legal protection system. On this basis, it explores, from three levels—legislative model, institutional design, and implementation mechanisms—the reference value of the EU experience for China, and proposes pathway recommendations for an AI legal system with Chinese characteristics.