ABSTRACT
Generative AI systems, including Generative Adversarial Networks (GANs), transformer-based Large Language Models (LLMs), and diffusion models, can now create literary, musical, and visual works comparable to human creative output. Since its creation in the Statute of Anne 1710, copyright law has been based on the assumption of human authorship as the source and recipient of creative rights, and there is a lack of legal clarity in the area of authorship, ownership, and copyright eligibility of works created by artificial intelligence systems. This study compares the approaches of India, the United Kingdom, the United States, and the European Union. The United Kingdom has a statutory provision for computer-generated works under section 9(3) of the Copyright, Designs and Patents Act 1988. The United States has upheld a stringent human authorship criterion based on constitutional text, recently affirmed in Thaler v Perlmutter (2025). The Indian Copyright Act 1957 has a fledgling provision for computer-generated works in section 2(y), but judicial interpretation has been minimal. The European Union is overlaying its AI regulatory framework on its prior copyright framework. By comparing the statutes, judicial rulings, and policy considerations of these four jurisdictions, this study finds that all four hold the view that an AI system cannot be a copyright author, but none has yet developed a comprehensive and operationally robust framework beyond that point of consensus. The most urgent gap is the absence of a clear, practical test for when a human has exercised sufficient creative control over an AI-assisted work to warrant copyright protection. The study proposes codification of the human authorship requirement, a workable human control test, training data transparency, sui generis protection for AI outputs, and active international harmonisation through WIPO.
1. INTRODUCTION
1.1 BACKGROUND OF THE STUDY
The dynamic emergence of the artificial intelligence technologies in the past ten years has become the main factor that has transformed the field of creative production. Nowadays, generative AI systems, including Generative Adversarial Networks (GANs) and transformer-based Large Language Models (LLMs) and diffusion models, can now create literary, musical, and visual works that are similar to and even superior to what human beings can create. The development of systems like the GPT lines of OpenAI, Google Gemini, Stable Diffusion, and Midjourney has introduced a brand-new form of creative output in which the human artist, writer or composer may be limited to providing a text prompt or a high-level instruction. The consequences of this paradigm shift on the law of copyright are far-reaching because since its creation in the Statute of Anne 1710 the law has been based on the assumption of human authorship as the source and recipient of creative rights.
Each of the four key intellectual property jurisdictions explored in this dissertation, i.e., India, the United Kingdom, the United States of America, and the European Union has approached this challenge with a certain level of legislative foresight, judicial creativity, and policy coherence. The United Kingdom has a unique status in that it has a statutory provision of computer-generated works, which are provided under the section 9(3) of the Copyright, Designs and Patents Act 1988 (CDPA). The United States has upheld a stringent human authorship criterion that is based on the constitutional text, which the DC Circuit Court recently affirmed in Thaler v Perlmutter (2025).
The Indian copyright Act 1957 has a fledgling provision of computer-generated works in section 2(y) of the Indian copyright act, but judicial interpretation has been minimal and the laws are not keeping up with technological advances. The European Union, long active in harmonisation of copyright standards by directives and the jurisprudence of the Court of Justice of the European Union (CJEU) is currently undertaking a complicated exercise of overlaying its AI regulatory framework (the EU AI Act 2024) on its prior copyright framework, especially the text and data mining exceptions of the Digital Single Market (DSM) Directive 2019.
The present dissertation is a comparative legal study of these four jurisdictions that are done systematically. It aims not only to determine the differences in the statutory text but also the policy preferences, tradition of doctrine, and judicial philosophy that have influenced each jurisdiction in its reaction to AI-generated creativity.
.jpeg)
