The Principle of Transparency in AI Regulation: A Comparative Legal Analysis of the European and Chinese Approaches

Kharitonova, Yulia and Lendvai, Gergely Ferenc (2025) The Principle of Transparency in AI Regulation: A Comparative Legal Analysis of the European and Chinese Approaches. KUTAFIN LAW REVIEW, 12 (4). pp. 801-828. ISSN 2713-0525

Preview

Text 723-1349-1-SM.pdf - Published Version Available under License Creative Commons Attribution Non-commercial. Download (279kB) | Preview

Abstract

The paper examines the principle of transparency in artificial intelligence (AI) regulations in two legal frameworks, namely, the European Union (EU) and China. The study aims to explore how transparency, a key principle for ensuring accountability and fostering trust in AI technologies, is regulated in these two distinct geopolitical environments. Using a comparative legal analysis approach, the paper reviews primary legal documents, scholarly literature, and expert analyses to identify commonalities and divergences in AI transparency regulations. The findings indicate that the EU’s AI Act emphasizes a risk-based approach, categorizing AI systems into high-risk, limited-risk, and minimal-risk categories, with stringent transparency requirements for high-risk systems. These requirements include comprehensive documentation, human oversight, and explainability to ensure that AI systems operate within ethical and legal boundaries. However, the AI Act also holds challenges, particularly for smaller enterprises, in meeting these transparency demands, as well as the technical difficulties in achieving transparency in complex AI models. In contrast, China’s regulatory framework, while similarly focused on transparency, integrates socialist moral and ethical values. The Chinese approach categorizes AI systems based on risk and emphasizes the interpretability and explainability of AI systems to ensure compliance with state-sanctioned moral principles. The findings suggest that while both the EU and China recognize the importance of transparency, their regulatory frameworks reflect broader cultural and political differences. The study concludes that achieving harmonized global AI transparency standards will require ongoing technological innovation, legal refinement, and international cooperation.

Actions (login required)

Edit Item