Artificial Intelligence and the Attribution of Criminal Liability

The advent of artificial intelligence presents profound challenges to the traditional framework of criminal law, particularly in the attribution of liability. While weak AI systems can be integrated into existing doctrines through the liability of developers, operators, or users, the realization of strong AI would raise fundamental questions concerning agency, autonomy, and the very foundations of responsibility. Current debates over free will, capacity for responsibility, and the so-called “responsibility gap” demonstrate that criminal law is not yet prepared to ascribe independent liability to AI systems. Comparative analysis reveals divergent approaches. The European Union has adopted a comprehensive regulatory framework through the AI Act, emphasizing risk classification, transparency, and accountability. The United States relies on sector-specific and case-driven responses. Korea has enacted the Framework Act on AI, establishing general principles of human-centered and trustworthy AI. By contrast, China has concentrated on adjacent domains such as data security, personal information protection, algorithm governance, and platform responsibility, while leaving questions of criminal liability largely within the broader framework of administrative enforcement and state control. Against this backdrop, the implications for Korea are clear. First, existing doctrines—such as negligence liability, the guarantor’s duty, and foreseeability—must be systematically reinterpreted in light of the particularities of AI-related harms. Second, new legislative measures are urgently required. In particular, provisions on algorithmic discrimination, explainability, transparency of AI decision-making, and cross-border liability should be prioritized. Third, international cooperation must be strengthened. Given the transnational nature of AI systems, Korea should actively engage in global norm-setting through institutions such as the OECD, the EU, and collaborative dialogues with the United States. Ultimately, a balanced legal framework that harmonizes technological innovation with the fundamental principles of criminal law must be developed. While artificial intelligence may transform human society, it cannot alter the legal premise that liability presupposes human capacity and judgment. Korean legislators and scholars must therefore respond both cautiously and proactively, ensuring that justice, fairness, and responsibility remain intact in the age of artificial intelligence.