Civil Law Status of Artificial Intelligence Systems: Legal Personality or Legal Regime of a Complex Object?

The rapid advancement of artificial intelligence (AI) technologies poses fundamental challenges to traditional civil law frameworks worldwide. This article examines the dichotomy between two principal approaches to determining the civil law status of AI systems: recognizing AI as a subject of law with electronic personhood versus treating AI as a complex object within the existing property law framework. Through comprehensive comparative legal analysis of international approaches, particularly the European Union's regulatory initiatives, developments in the United States, and emerging practices in Asia and the Middle East, this study evaluates the applicability of each model to the legal system of the Republic of Kazakhstan. The methodology combines doctrinal legal analysis with comparative law methods, examining primary legal sources including the EU AI Act (Regulation 2024/1689), the EU Product Liability Directive (2024/2853), Kazakhstan's Law «On Artificial Intelligence» (No. 230-VIII, 2025), and relevant case law from multiple jurisdictions. The empirical dimension includes analysis of documented AI-related incidents in transportation, healthcare, and financial services sectors that illustrate the practical challenges of applying traditional liability frameworks to autonomous systems. The research concludes that granting full legal personality to AI systems is premature and creates potential legal loopholes that could shield human actors from responsibility. The Saudi Arabian grant of citizenship to the robot Sophia in 2017, while symbolic, demonstrates the conceptual confusion that can arise from treating AI as a legal subject. The most viable model for Kazakhstan's legislation is recognizing AI as a complex object with a special liability regime, building on the existing concept of «sources of increased danger» in Article 931 of the Civil Code. The article proposes specific amendments to the Civil Code of the Republic of Kazakhstan, including: (1) explicit classification of high-risk AI systems as sources of increased danger; (2) establishment of joint and several liability among AI supply chain participants; (3) introduction of rebuttable presumptions regarding causation and defectiveness; (4) implementation of mandatory insurance requirements modeled on existing vehicle liability insurance; and (5) enhanced disclosure obligations to address information asymmetries in AI litigation. These recommendations aim to create a balanced framework that protects victims of AI-caused harm while fostering responsible innovation in Kazakhstan's emerging AI sector.