Artificial intelligence as a participant in civil relations: the transformation of law

Abstract

Artificial Intelligence (hereinafter referred to as AI) carriers are active participants in civil relations. Therefore, questions of legal nature of AI units, their ability to possess rights and bear responsibilities, exercise rights and perform responsibilities, questions of liability limits for improper exercise of rights and performance of responsibilities by AI units are most relevant. For law, the defining characteristic of any participant in civil relations is its role in the exercise of rights and the performance of duties. The digital actor must also be viewed from this perspective. One of the central debates in this area is the possibility of conferring legal capacity on digital actors. This position has found supporters and opponents both in domestic science and abroad. At the same time it is necessary to take into account the specifics of will, consciousness and other anthropomorphic qualities of AI carriers. Complex neural networks have similar anthropomorphic qualities, the nature of which is a digital algorithm. AI is capable of expressing its will in accordance with the interest, goals and motives set by software algorithms, i.e. it is technological in nature. Obviously, an AI must be endowed with legal capacity and be responsible for its actions, otherwise we cannot use the full potential of an AI as a technology (person), and the actions of such a participant in legal relations cannot be recognised as legally significant. According to its purpose and registration an AI medium can be endowed with different legal personality: to be an agent - a subject with a narrow special legal capacity within certain functions, to participate in civil relations as a full subject, or to act as an object of legal relations at all. This legal capacity is conditioned by the robot having a certain set of qualities, such as the level of self-awareness, autonomy, complexity of neural networks, and competences. It is necessary to create a classification of AI forms by legal personality, by competences, which are certified through state registration. Legal statuses of different types of AI should be defined, including a new specific legal status of limited digital legal capacity. It is necessary to consider the legal personality of AI in relation to its exercise of rights, principles of law and legal content. It is necessary to create for AI varieties a certain mechanism of responsibility conditioned by the legal status assigned to it at the state registration. With respect to AI carriers the nature of law execution changes - subjective expression of will disappears, the execution becomes objective, unbiased. AI software must incorporate a system of principles for the exercise of rights and the performance of duties. In this way, the AI will be certified to a specific scope of powers, be competence-oriented and have a specific legal capacity.