Generative Artificial Intelligence (Generative AI) has transformed the creation of literary, artistic, musical, dramatic and audiovisual works. Unlike conventional computer-assisted creativity, Generative AI systems independently produce sophisticated outputs after processing vast datasets and responding to human prompts. This development challenges copyright law at its foundation, because traditional regimes are constructed around a human author, originality, ownership and exclusive rights. The central legal question is whether an output generated substantially or entirely by an AI system attracts copyright and, if so, who is to be recognised as its author and owner. The difficulty intensifies where the user supplies only a prompt, where the user undertakes extensive creative selection and modification, or where the system autonomously produces the final expression. The Copyright Act, 1957 contains no dedicated framework for Generative AI, although Section 2(d)(vi) on computer-generated works, Section 13 on subsistence, Section 17 on first ownership and Sections 18 and 19 on assignment supply the governing principles. The order of the Registrar of Copyrights of 31 August 2026 in the DABUS matter holds that an artificial intelligence system, having no legal personality, cannot be the statutory author, although its output may satisfy originality. The United States requires human authorship, the United Kingdom protects computer-generated works by statute, and the European Union regulates providers of general-purpose AI models through copyright and transparency obligations. This paper examines the Indian framework and the relevant comparative developments, the first Indian judicial ruling on model training, and the contemporary controversies concerning authorship, ownership, training data, infringement and accountability. It argues that neither unrestricted copyright for autonomous AI outputs nor the complete exclusion of AI-assisted works is adequate. A human-centred but technology-neutral framework, supported by disclosure requirements, provenance mechanisms, licensing systems and legislative clarification, is necessary to balance technological innovation against the legitimate interests of human creators.