ARTIFICIAL INTELLIGENCE: SUBJECT OR OBJECT OF CIVIL LEGAL RELATIONS?

The article is devoted to the analysis of the legal status of artificial intelligence (AI) in civil legal relations, which is one of the most debated issues in the legal field. It examines modern approaches to determining AI’s legal subjectivity, as well as the problem of liability for its autonomous actions. The article presents a comprehensive analysis of existing theoretical approaches to defining the legal status of AI, particularly considering the object-based, subject-based, and hybrid approaches used in legal doctrine. Special attention is given to the issue of liability for AI’s autonomous actions, which is a critically important aspect in the context of civil legal relations. It is determined that Ukraine’s national legislation currently lacks clear provisions regulating the legal subjectivity of AI or the distribution of legal responsibility for damage caused by autonomous systems. The study analyzes approaches applied in international practice, particularly within the European Union (AI Act, 2021) and the United States (Algorithmic Accountability Act, 2022). Based on the analysis of judicial practice, including cases related to the use of autonomous vehicles and algorithmic decision-making systems (such as the Uber case (2018) and Amazon’s discriminatory hiring algorithm case), the article identifies key legal challenges associated with AI’s legal status. The concept of “digital personality,” proposed in the works of Western scholars, is explored as a potential legal classification for AI. The article also formulates proposals for improving Ukraine’s regulatory framework in the field of AI governance. It suggests the development of a special law, “On Artificial Intelligence,” which would define levels of AI system autonomy, mechanisms for distributing liability among developers, users, and owners, as well as a system of state oversight for high-risk AI technologies. Special emphasis is placed on the need to harmonize Ukrainian legislation with EU standards in the field of AI regulation. The study concludes that the most appropriate approach to the legal regulation of artificial intelligence is the hybrid model, which combines AI autonomy with elements of legal oversight by humans.