On some legal approaches to determining the legal personality of artificial intellectual systems

The issues of determining the place of artificial intellectual systems in the structure of new legal relations determined by the level of technological development are one of the most urgent tasks of modern Russian law today. The most important problem, determined by the requirements of law enforcement and the need to comply with the requirements of building of the rule-of-law state, is the study of fundamental positions regarding the conditions and procedure for bringing subjects to legal responsibility for torts involving artificial intelligence (AI). Purpose: to analyze approaches to the definition and content of legal personality, in particular, the possible legal personality of artificial intellectual systems, with an emphasis on the rules for imposing liability for torts involving them. Methods include dialectical and metaphysical methods of cognition (as the philosophical basis of the work performed); general scientific methods (analysis, synthesis, abstraction, hypotheses, etc.); as well as special scientific (comparative-legal, legal-dogmatic, cybernetic, interpretation) methods of scientific cognition. Results: the study, for the purpose of determining the legal personality of AI, justifies the greatest significance of its classification into: weak narrow-purpose AI (Narrow AI), strong general-purpose AI (General AI), and super-strong intelligence (Super AI). Based on the materials of scientific sources, the concept of «partial legal capacity» is formulated and the possibility of its application to strong and super-strong AI is scientifically justified. Fundamental theoretical positions (General rule and exceptions) regarding the conditions and procedure for bringing subjects to legal responsibility for torts involving AI are presented.