Copyright Facing the Challenges of Generative Artificial Intelligence: Judicial Practice and Legislative Strategies in India, the United States and the European Union

Objective: to conduct a comparative analysis of the judicial interpretation of the fair dealing and fair use doctrines in the copyright law systems of India, the United States and the European Union in the context of the challenges posed by the development of generative artificial intelligence and blockchain technologies. Methods: the work uses a set of scientific methods, including a comparative legal analysis of the legislation of three jurisdictions, a systematic analysis of judicial practice in India, a dogmatic method of interpreting regulations, as well as a structural and functional approach to the study of legal institutions. Special attention was paid to over sixty years of Indian judicial practice in applying the fair dealing doctrine, to the American fair use doctrine with its four-factor test, and to the European system of legislative exceptions in text and data mining. The research methodology includes a historical and legal method for identifying evolutionary trends in the judicial interpretation of copyright exceptions, a formal legal method for analyzing the normative content of legal institutions, and a legal modeling method for developing recommendations to improve legislation for regulation of generative artificial intelligence and blockchain technologies. Results: the study convincingly demonstrates the structural inconsistency of the Indian closed-list system of copyright exclusions for regulating generative artificial intelligence and blockchain technologies. It was established that the Indian fair dealing doctrine is characterized by five fundamental limitations: excessively literal interpretation of the legislative text, lack of a transformative use concept, inability to adapt to digital formats, legal gap in the regulation of the artificial intelligence outputs, and significantly limited application. A comparative analysis revealed that the American system reaches structural limits when regulating the large-scale use of data, whereas the European model covers the data input but not the commercialization of artificial intelligence outputs. Scientific novelty: the research presents a comprehensive comparative legal analysis of the application of the fair dealing and fair use doctrines to generative artificial intelligence and blockchain technologies. The study systematizes more than sixty years of judicial practice in three legal systems, which allowed identifying the structural limitations of both open and closed models of copyright exceptions and justifying the need to comprehensively regulate full cycle of the creation and commercialization of artificial intelligence content. Practical significance: the results can be used to develop national strategies for regulating artificial intelligence; reform the system of copyright exceptions; introduce technologically neutral standards for text and data mining; create disclosure mechanisms for training datasets and registers of copyright holders’ opt-outs; and modernize the system of collective rights management using blockchain.