LEGAL LIABILITY FOR DECISIONS MADE BY ARTIFICIAL INTELLIGENCE

This article analyzes theoretical and practical problems of legal liability for decisions made by, or with the participation of, artificial intelligence (AI) systems. The study examines the notion of the “responsibility gap”, the legal status of AI, and the allocation of liability among developers, providers, deployers and final human decision-makers from civil, administrative and criminal law perspectives. The Law of the Republic of Uzbekistan of 21 January 2026 and the national AI ethics rules are compared with the EU Artificial Intelligence Act and the new EU Product Liability Directive. The research concludes that granting legal personality to AI is inappropriate and proposes a multi-tier liability model together with scientific and practical recommendations for its implementation.