Comparative analysis of national models of legal regulation of artificial intelligence in the context of ensuring the safety of human rights and innovative development

Abstract

By 2026 national legal orders had developed heterogeneous approaches to the regulation of artificial intelligence, and the divergences between them acquired a stable character that affects the very architecture of regulation. Regulatory policy is structured around the tension between ensuring the safety of human rights and stimulating technological development, and the choice of the dominant objective predetermines the type of legal instruments applied. The study provides a comparative legal analysis of four national models: the horizontal risk-based regime of the European Union, the decentralized sectoral approach of the United States of America, the vertical administrative regime of the People's Republic of China, and the predominantly strategic regime of the Russian Federation. The comparison is carried out according to the criteria of the priority of the regulatory objective, the type of legal instruments, and the design of the risk architecture. It is established that the European model achieves the protection of rights at the cost of a high regulatory burden, the American one shifts the balance towards innovative development by refusing a single federal act, the Chinese one combines strict content control with the encouragement of sectoral adoption, and the Russian one keeps rights and development in the strategic plane amid a shortage of directly applicable acts. The juxtaposition of the regimes reveals that the opposition between the safety of rights and innovative development is not absolute but instrumental: different models reach a compromise between them not by abandoning one of the goals but by distributing the regulatory burden across the levels of the normative system and the stages of the technology life cycle.