AUTONOMY AS A CRITERION OF LEGAL PERSONALITY OF ARTIFICIAL INTELLIGENCE AND ITS APPLICATION IN THE FIELD OF INTELLECTUAL RIGHTS

Abstract

The article examines the concept of autonomy in relation to artificial intelligence (AI) in the context of regulatory regulation and the positions of researchers referring to autonomy as an integral characteristic of AI. Autonomy, the ability to become autonomous, the ability to act autonomously, and so on are the criteria that the authors traditionally give to AI. Only some researchers do not use terminological constructions based autonomy in relation to AI definitions. At the same time a number of researchers, guided by the criterion of autonomy, draw conclusions about the need to endow AI with a special status, which, perhaps, will also determine the legal responsibility of AI. As a result, the science did not form a unified understanding of autonomy, and began to consider it as a matter of course when describing AI. But the inability of AI to have legal personality and act as an author is based exactly on the absence of the real autonomy. The concept of autonomy is also considered in relation to the criteria that guided the Patent Office of South Africa and the courts of China and Australia in 2021 when making decisions on granting AI the rights of the author / inventor (in Australia, the decision was overturned by the court of appeal in 2022). The most detailed decision was justified by the Federal Court of Australia, used a broader view of the concept of "inventor". To this it would be logical to add a broader view of the concept of "autonomy", in the application of which it was hardly possible to assert that it was AI, without "additional human participation» that could become an author or even exist.