Galina V. Bestolkova

This article involves a comparative legal analysis of three AI regulatory models: EU, USA, China. The given research’s relevance is determined by AI technologies’ rapid development and regulatory mechanisms’ significant lag behind technological progress. The research’s purpose is to identify conceptual differences and to reveal common trends in AI regulatory process within the three jurisdictions (EU, USA, China). The methodological basis of the research is three jurisdiction documents’ comparative legal analysis based on six criteria: regulation type, AI technologies’ classification, requirements for developers and providers, enforcement methods, system of fines, extraterritorial effect. The given research’s empirical basis consists of three jurisdictions’ regulatory documents governing AI technologies’ development. The study’s results demonstrate that European Union applies a risk-based approach with strict fines, United States use a flexible sectoral model with voluntary standards, China creates a state-centric system with mandatory algorithm registration. The given study’s results indicate that common problem across all three AI regulatory models is autonomous AI technologies’ comprehensive regulation lack. The research also suggests that extraterritorial scope of each AI regulatory model becomes a key factor in the competition to shape global AI regulatory architecture. The given research’s practical significance provides possibilities for adapting the most effective elements of each model to improve national AI development strategies: European Union’s risk classification, United States’ voluntary framework standards, and China’s mandatory algorithm registration. This approach can balance citizen rights’ protection, national security, and technological development, while avoiding considered models’ extremes.