Legal status of neural: through the prism of judicial practice

Abstract

The article analyzes the practical side of understanding neural networks in the legal aspect on the basis of case law of the USA, UK, Australia, EU and China and foreign legislation, in particular, the Copyright, Designs and Patents Act 1988; the Digital Millennium Copyright Act; the Indian Copyright Act; the Irish Copyright and Related Rights Act; the New Zealand Copyright Act, with practical examples.
 The following main aspects of neural networks functioning are mentioned: copyright subjects; the place of the object generated by a neural network in the legal dimension; neural network training materials, ways of obtaining them, and conflict issues.
 The current case law on neural networks and objects generated by them is divided into the following thematic groups: disputes concerning the determination of the copyright holder, establishment of legal personality and aimed at confirming the universal rule regarding an individual as an author; cases focusing on the possibility of copyright protection of specific objects generated by a particular neural network; disputes concerning the formation of a legal dimension for neural network training materials.
 Modeled groups of situations that may be reflected in different circumstances of the functioning of a neural network and artificial intelligence. The author emphasizes the legally significant component of a neural network. Attention is drawn to the doctrine of fair use and the concept of «temporary acts of reproduction», together with their sources and main features. Potentially negative consequences of copyright infringement by neural networks are also outlined.
 As conclusions, identified gaps in the legal regulation of neural networks and objects generated by them, which is reflected, in particular, in the lack of practical significance of a number of legal norms and doctrines in the context of dispute resolution and the existence of contradictions in the ideas regarding the criteria of originality and important nuances in cases of establishing or denying copyright infringement within the process of training neural networks on specific materials. The author emphasizes the need for further research in the field of law and technology.