AI-Generated Content and the Copyright Act, 1957 — Is There a Gap in Indian Law?

Introduction

The rapid advancement of Artificial Intelligence has fundamentally disrupted the creative ecosystem. Tools such as ChatGPT, DALL-E, Midjourney, and Suno AI are today capable of generating text, images, music, poetry, legal briefs, and cinematic script outputs that are, in many cases, indistinguishable from human-created works. These AI systems do not merely assist human creativity; they independently generate original creative output with minimal or no human authorial input.

This technological reality has collided with a legal framework that was never designed to anticipate it. The Copyright Act, 1957 - India's primary legislation governing intellectual property in creative works, was enacted in an era of typewriters and hand-drawn illustrations. Its foundational assumption is singular and unambiguous: only a human being can be an author. Every provision of the Act - from the definition of authorship to the calculation of copyright duration presupposes a living, breathing human creator.

The central question that emerges from this collision is both legally urgent and commercially consequential: When an AI system independently generates a creative work, who, if anyone, owns the copyright? Is it the developer who built the AI? The user who provided the prompt? The company that deployed the system? Or does the work fall into the public domain belonging to everyone and no one simultaneously?

This article examines the existing provisions of the Copyright Act, 1957, analyses the judicial and regulatory responses to AI-generated content in India and comparable jurisdictions, and argues that India faces a significant and commercially damaging legislative gap that demands urgent parliamentary attention.

The Foundation - What the Copyright Act, 1957, Actually Says

The Copyright Act, 1957 defines an "author" under Section 2(d) with reference to specific categories of works. For literary, dramatic, musical, and artistic works, the author is the person who creates the work. For cinematograph films, the producer. For sound recordings, the producer. For computer-generated works, the person who causes the work to be created.

This last category is the most legally significant for AI. Section 2(d)(vi) - inserted by the Copyright (Amendment) Act, 1994 - specifically addresses computer-generated works and vests authorship in "the person who causes the work to be created." At first glance, this provision appears to resolve the AI authorship question: the user who prompts the AI, or the developer who built it, is the "person who causes the work to be created."

However, this interpretation is fraught with complexity. The 1994 amendment was enacted to address works generated by basic computational tools, spreadsheet outputs, word-processed documents, and algorithmically arranged databases. The legislature of 1994 could not have contemplated a system like Midjourney, which generates a photorealistic painting from a three-word prompt with no meaningful human creative contribution to the artistic output itself. The question is not merely who pressed the button - but whether pressing a button constitutes the kind of human creative authorship that copyright law was designed to protect and reward.

The Originality Requirement - The Core Legal Problem

Copyright protection in India, as in most jurisdictions, requires that a work be original. While the Copyright Act does not explicitly define originality, Indian courts have adopted the skill-and-judgment standard, requiring that a work reflect the author's own intellectual effort, skill, and creative expression.

The Supreme Court of India affirmed this in Eastern Book Company v. D.B. Modak (2008), holding that a work must reflect the author's own intellectual creation to qualify for copyright protection mere mechanical reproduction does not suffice.

This originality standard creates a direct and irresolvable problem for AI-generated content. An AI system does not exercise skill and judgment in the human sense. It applies statistical pattern recognition to training data - producing outputs that reflect the patterns of millions of human-created works it was trained on, not any original intellectual effort of its own. The AI has no creative intent. No aesthetic vision. No personal expression. It executes a probabilistic algorithm. If originality requires human intellectual creation, AI-generated works are not original. And if they are not original, they are not protected by copyright.

The Human Authorship Requirement - A Constitutional Dimension

Beyond the originality standard, copyright law globally has consistently held that only a human being can be an author. In the United States, the Copyright Office has issued clear guidance in the Thaler v. Vidal (2022) line of cases and subsequent administrative decisions that works created autonomously by AI without human creative expression are not eligible for copyright registration. The US District Court in Thaler v. Perlmutter (2023) expressly held that copyright requires human authorship and rejected the registration of a work created entirely by an AI system called DABUS.

In India, there is no equivalent judicial or regulatory guidance. The Copyright Office has not issued any policy statement on AI-generated works. The Intellectual Property India portal continues to process applications on the basis of existing law which assumes human authorship. The absence of guidance has created a vacuum in which commercially valuable AI-generated works are being created daily by Indian businesses, music composed by AI for advertising campaigns, articles written by AI for digital publications, images generated by AI for product catalogues with no clarity on who owns these works or whether they are protected at all.

The Ownership Problem - Multiple Competing Claims

Even if one accepts that Section 2(d)(vi) of the Copyright Act provides a route to copyright protection for AI-generated works the question of who exactly is the "person who causes the work to be created" remains deeply contested.

The company that built the AI - OpenAI, Stability AI, Midjourney Inc. invested billions of dollars in research, trained the model on vast datasets, and created the technical infrastructure that enables the generation. Their terms of service typically assign ownership of outputs to the user but the legal enforceability of such contractual arrangements against third parties and under Indian law is entirely untested.

Alternatively, the person who typed the prompt arguably "caused the work to be created" within the meaning of Section 2(d)(vi). But if the prompt is merely three words - "paint a sunset" - does this constitute sufficient creative contribution to merit authorship? If neither the AI, nor the developer, nor the user satisfies the originality and authorship requirements the work falls into the public domain - unprotected, free for anyone to use, copy, and commercialise. This may be the legally correct outcome under existing Indian law - but it is commercially catastrophic for businesses that have invested in AI-generated content.

The Training Data Problem - A Parallel Copyright Issue

The AI authorship question is compounded by a parallel and equally urgent issue: the legality of training AI systems on copyrighted works. Modern generative AI systems are trained on enormous datasets scraped from the internet - datasets that inevitably contain millions of copyrighted works, including books, articles, paintings, photographs, and music, reproduced without the consent of their human authors.

Indian copyright law does not provide a specific fair use exception for AI training. Section 52 of the Copyright Act, 1957 - which lists acts not constituting infringement does not mention AI training or machine learning. Whether training an AI on copyrighted works constitutes infringement under Indian law is entirely unresolved. This question is currently being litigated in the United States, with cases like Andersen v. Stability AI and The New York Times v. OpenAI raising fundamental questions about whether AI training constitutes copyright infringement. India will inevitably face the same litigation, but without the benefit of even the preliminary judicial framework that is slowly emerging in American courts.

Comparative Perspectives - What Other Jurisdictions Are Doing

The United Kingdom's Copyright, Designs and Patents Act, 1988 contains a provision - Section 9(3) that specifically addresses computer-generated works, vesting copyright in "the person by whom the arrangements necessary for the creation of the work are undertaken." This is the most forward-thinking legislative provision currently in force anywhere in the world on this question and India's 1994 amendment to Section 2(d) was directly inspired by it.

The European Union has not yet adopted specific legislation on AI copyright. The EU AI Act, 2024 the world's first comprehensive AI regulatory framework, addresses AI safety and transparency but does not resolve copyright ownership of AI outputs. China's Beijing Internet Court in Li Yunkai v. Liu Yuanchun (2023) became the first court in the world to hold that AI-generated images can be protected by copyright provided the human user exercises sufficient creative control over the generation process. This decision is highly significant though it remains a first-instance decision with limited precedential value.

The Legislative Gap - India's Urgent Problem

India's Copyright Act, 1957 has not been amended since 2012. The 2012 amendment addressed digital piracy and online copyright enforcement; it did not contemplate generative AI, which was then in its infancy. The result is a legislative framework that is dangerously out of step with technological reality. Indian businesses creating AI-generated content in a rapidly growing and commercially significant sector have no legal certainty about whether their content is protected, who owns it, or whether the AI systems they use have infringed the copyrights of the works they were trained on.

The Parliamentary Standing Committee on Commerce has noted the need to review India's intellectual property framework in light of emerging technologies but no specific bill addressing AI and copyright has been introduced in Parliament. The National IPR Policy, 2016 India's governing framework for intellectual property does not mention artificial intelligence at all. In 2016, this was understandable. In 2025, it is a serious legislative failure.

Recommendations

India requires, as a matter of urgency, the following legislative and regulatory interventions:

• An amendment to Section 2(d) of the Copyright Act, 1957 - clarifying the authorship of AI-generated works and specifying whether and under what conditions such works qualify for copyright protection.

• A fair use exception for AI training - modelled on the text and data mining exceptions enacted in the UK and EU - to provide legal certainty for Indian AI developers and researchers.

• A mandatory disclosure requirement- requiring creators to disclose when a work was generated wholly or substantially by AI - to protect human authors and consumers.

• A policy statement from the Copyright Office of India - providing interim guidance to applicants and businesses pending legislative amendment.

Conclusion

The Copyright Act, 1957 was built for a world of human creativity. Artificial intelligence has created a new world - one in which machines generate creative works at an industrial scale, with no human hand guiding the brush or the pen. Indian copyright law has not kept pace with this transformation.

The gap is not merely technical or academic. It is commercially damaging, constitutionally significant, and legally urgent. Businesses investing in AI-generated content deserve legal certainty. Human authors whose works trained the AI deserve protection. Consumers encountering AI-generated content deserve transparency.

India has an opportunity and an obligation to lead the global conversation on AI and intellectual property. A country that aspires to be a five trillion dollar digital economy and a global hub for artificial intelligence cannot afford a copyright law that belongs to 1957. The legislature must act and it must act now.

Author: Vaishnavi M, in case of any queries please contact/write back to us via email to chhavi@khuranaandkhurana.com or at  Khurana & Khurana, Advocates and IP Attorney.

References

1. Copyright Act, 1957 (India) — Sections 2(d), 13, 17, 52

2. Copyright (Amendment) Act, 1994 (India)

3. Copyright (Amendment) Act, 2012 (India)Eastern Book Company v. D.B. Modak, (2008) 1 SCC 1 (Supreme Court of India)

4. Eastern Book Company v. D.B. Modak, (2008) 1 SCC 1 (Supreme Court of India)

5. Thaler v. Vidal, 43 F.4th 1207 (Fed. Cir. 2022) (United States Court of Appeals)

6. Thaler v. Perlmutter, Case No. 22-cv-1564 (D.D.C. 2023) (US District Court)

7. Li Yunkai v. Liu Yuanchun, Beijing Internet Court (2023) (People's Republic of China)

8. Andersen v. Stability AI Ltd., Case No. 3:23-cv-00201 (N.D. Cal. 2023)

9. The New York Times Company v. Microsoft Corporation, Case No. 1:23-cv-11195 (S.D.N.Y. 2023)

10. UK Copyright, Designs and Patents Act, 1988 — Section 9(3)

11. EU AI Act, 2024 — Regulation (EU) 2024/1689 of the European Parliament and of the Council

12. National IPR Policy, 2016 — Department for Promotion of Industry and Internal Trade, Government of India

13. World Intellectual Property Organization (WIPO), 'Revised Issues Paper on Intellectual Property Policy and Artificial Intelligence' (WIPO/IP/AI/2/GE/20/1 REV)

14. Digital Personal Data Protection Act, 2023 (India)