This piece of article is authored By:Asha Merotha 3rd Year law student at Symbiosis Law School, Pune &Sejal verlani 4th year law student at Institute of law , Nirma University.
I. Introduction
The rapid proliferation of artificial intelligence technologies has fundamentally challenged the foundational premises of intellectual property law in India. The Copyright Act, 1957—which remains the primary statutory instrument governing authorship and ownership in India—was conceived in an era when copyright inhered exclusively in human creative endeavor. Today, generative AI platforms such as ChatGPT, DALL-E, and other autonomous systems create original literary, artistic, and audiovisual works with minimal human intervention. This technological advance has exposed critical lacunae in India’s copyright framework regarding authorship determination, ownership attribution, and the rights of content creators whose works are utilized for AI training without authorization. The emergence of disputes such as Asian News International v. OpenAI Inc. (pending before the Delhi High Court) and the contested registration of RAGHAV—an artificial intelligence system claiming co-authorship of visual artwork—demonstrates that Indian courts must urgently reconcile nineteenth-century statutory language with twenty-first-century technological reality. This article examines the statutory framework governing AI-generated content under Indian law, analyzes seminal case law establishing judicial precedent, and identifies necessary policy reforms to protect both creators and foster innovation.
II. Statutory Framework and Authorship Under Indian Law
The Copyright Act, 1957 provides the foundational legal architecture for copyright protection in India, yet offers no explicit guidance on artificial intelligence or machine-generated works. Section 2(d) defines the author as the originator of a literary, dramatic, musical, or artistic work, with distinctive provisions for cinematographic films (authored by the producer), sound recordings (authored by the producer), and computer-generated works authored by “the person who causes the work to be created.” This last provision—added through amendment—creates interpretive ambiguity when applied to fully autonomous AI systems. The statutory language “person who causes the work to be created” presupposes intentional human causation and creative agency. Courts have consistently rejected extending legal personhood to non-human entities absent explicit legislative mandate. In Eastern Book Company v. D.B. Modak, the Supreme Court established that originality in copyright derives from “skill and judgment”—requiring intellectual effort and creative contribution rather than mechanical application of computational tools. The Supreme Court held that merely using computer technology to manipulate existing material without independent creative contribution fails to meet the originality threshold necessary for copyright eligibility. This precedent, though predating contemporary AI, establishes that algorithmic processing alone cannot constitute authorship. The statutory framework thus confronts an unresolved tension: Section 2(d) extends authorship to “the person who causes” creation, yet extends no recognition to artificial entities that operate substantially autonomously from human direction.
III. Statutory Limitations and the Fair-Dealing Framework
Section 52 of the Copyright Act, 1957 permits limited non-infringing uses of copyrighted material under the doctrine of “fair dealing,” including uses for research, private study, criticism, review, and journalism. These provisions contemplate human analytical purposes and scholarly engagement. The statute does not explicitly address whether mechanistic AI training—which processes copyrighted material algorithmically through vectorization, storage, and pattern extraction—constitutes fair dealing or constitutes copyright infringement. The pending ANI v. OpenAI case directly addresses this question. Asian News International, an Indian news agency, filed suit in the Delhi High Court in November 2024 alleging that OpenAI used ANI’s copyrighted news articles without authorization to train ChatGPT’s large language model. ANI contends that despite content being publicly accessible online, OpenAI’s use of such material for AI training constitutes commercial exploitation of intellectual property without compensation. OpenAI’s defense rests on two propositions: first, that AI training on publicly available data constitutes industry standard practice globally and does not amount to copyright infringement; second, that publishers seeking protection should implement technological barriers such as paywalls or anti-scraping measures rather than placing compliance burdens on AI developers. The case raises three distinct infringement questions under Section 14 of the Act: whether storage of copyrighted material for training purposes constitutes prohibited reproduction (Section 14(a)(i)), whether use of stored material in generating responses constitutes prohibited adaptation (Section 14(a)(vi)), and whether commercial scale and profit-motive eliminate fair-dealing protection. The amicus curiae appointed to assist the Delhi High Court submitted that fair dealing protection for AI training data must be rigorously construed. Adarsh Ramanujan argued that copying content without permission—even once—constitutes infringement, and that OpenAI’s use cannot qualify as fair dealing since it operates neither as news agency, critic, nor reviewer as contemplated by Section 52. This positions fair-dealing doctrine in direct tension with practical AI development, as comprehensive training datasets are non-negotiable for algorithmic performance.
IV. Landmark Cases and Judicial Precedent on AI Authorship
The trajectory of Indian copyright jurisprudence on artificial intelligence authorship reveals judicial reluctance to extend authorship beyond human creators absent explicit statutory authorization. In R.G. Anand v. Delux Films, decided by the Supreme Court in 1978, the court established that copyright protects the expression of ideas rather than ideas themselves. This foundational principle creates interpretive space for determining what constitutes sufficient human “expression” in AI-assisted works. The holding suggests that if a human user provides creative direction through prompting or iterative refinement, thereby substantially shaping the final work’s expression, copyright protection may inhere in that human contribution—even if the AI executes the mechanical aspects of creation. However, purely autonomous AI outputs lacking human direction regarding expression remain unprotected.
The most direct judicial engagement with AI authorship in India appears in the RAGHAV case. Ankit Sahni applied for copyright registration identifying RAGHAV (Robust Artificially Intelligent Graphics and Art Visualizer) as co-author of “Suryast,” a digitally recreated image derived from a photograph of sunset composed in the style of Vincent van Gogh’s Starry Night. The Indian Copyright Office initially rejected the application listing RAGHAV as sole author, explicitly citing the lack of legal personhood for artificial entities. However, upon revision, the office granted copyright registration with Sahni designated as author and RAGHAV listed as co-author. This registration remained on file but was subsequently placed under withdrawal notice, with the Copyright Office demanding juridical clarification of RAGHAV’s status. The Delhi High Court has not yet issued final judgment on the merits; however, the progression of administrative decisions demonstrates that Indian copyright authorities have begun interrogating whether human-AI collaborative works merit protection. The case fundamentally challenges whether Section 2(d)’s reference to “the person who causes the work to be created” can encompass machine-assisted creation where human input, though substantial, does not extend to independent ideation or mechanical execution.
The DABUS case, adjudicated by the High Court of England and Wales, provides instructive comparative authority. In Stephen L Thaler v. Comptroller General of Patents, the appellant sought patent protection for inventions allegedly generated entirely by an AI system named DABUS (Device for the Autonomous Bootstrapping of Unified Sentience), listing DABUS as the inventor. The English High Court rejected this claim, holding that machines cannot qualify as patent inventors absent legal personhood. The court reasoned that patent law presupposes intentional inventive activity by conscious agents capable of understanding the technical field, identifying problems, and developing solutions through cognitive exercise. DABUS, lacking consciousness, intentionality, and cognitive understanding, could not satisfy this requirement. The judgment’s logic extends with considerable force to copyright authorship, as both intellectual property regimes presuppose conscious creative intent. This comparative analysis undermines any argument that Indian copyright law could extend authorship to fully autonomous AI systems through interpretive expansion without legislative amendment.
The New York Times v. OpenAI case, pending before the United States District Court for the Southern District of New York, establishes additional precedent regarding copyright infringement through AI training and output generation. The Times alleges that OpenAI and Microsoft used its proprietary journalistic content to train language models and that ChatGPT subsequently generated responses reproducing substantial excerpts from Times articles verbatim or near-verbatim, particularly through Microsoft’s “Browse with Bing” feature powered by ChatGPT. The complaint demonstrates that AI-generated responses attributed false information to the Times, including listings of dietary recommendations allegedly from the Times but absent from any published article. This case positions copyright infringement liability on two separate grounds: unauthorized copying for training and subsequent generation of output that reproduces substantial portions of protected expression. While the United States applies fair-use doctrine rather than India’s fair-dealing framework, the underlying copyright infringement claims address statutory provisions structurally similar to Indian law.
V. Legislative Lacunae and Policy Reform Imperatives
Indian copyright law contains no express provisions addressing AI authorship, machine-generated content ownership, or licensing regimes governing AI training data usage. This legislative vacuum creates three unresolved questions that demand urgent statutory clarification. First, can artificial intelligence systems qualify as authors or co-authors under current law, or does authorship remain exclusively reserved for human agents? Section 2(d)’s reference to “the person who causes the work to be created” remains indeterminate when applied to fully autonomous systems operating without human direction regarding specific creative outputs. Second, who holds ownership rights when human creativity inputs (such as detailed prompting) combine with AI execution to generate final creative works? The contribution-to-expression framework from R.G. Anand v. Delux Films offers guidance but requires explicit statutory codification to provide certainty for creators and technology developers. Third, what licensing or opt-out mechanisms should govern use of copyrighted works for AI model training? The ANI v. OpenAI litigation demonstrates that copyright owners experience substantial economic harm when training datasets are compiled without authorization or compensation. The Ministry of Commerce established an eight-member expert panel in April 2025 specifically tasked with examining copyright implications of artificial intelligence and formulating policy recommendations. This governmental recognition that existing law proves inadequate creates opportunity for comprehensive legislative reform.
India’s National IPR Policy (2016) emphasizes copyright as instrumental to promoting cultural creation and commerce. The policy articulates commitment to “awareness, promotion, and respect for intellectual property rights” and prioritizes “enforcement of rights and prevention of violations.” Yet the policy predates contemporary AI by years and contains no guidance for AI-era copyright challenges. Comparative international frameworks demonstrate varied policy responses. The United Kingdom’s Copyright, Designs and Patents Act, 1988 explicitly recognizes computer-generated works under Section 9(3), designating “the person who made the necessary arrangements” as author. This statutory clarity contrasts sharply with Indian law’s silence. The European Union’s emerging AI regulation requires authorization from copyright holders before using protected works for AI model training and mandates transparency in AI model development. These comparative models suggest that India must choose between (i) expansive judicial interpretation extending existing copyright categories to accommodate AI contexts, or (ii) comprehensive statutory amendment establishing AI-specific intellectual property rights.
VI. Conclusion: Toward a Coherent Legal Framework
India stands at a regulatory crossroads regarding artificial intelligence and copyright protection. The current statutory framework—predicated entirely on human authorship and conscious creative intent—provides inadequate guidance for AI-era challenges. Judicial decisions in RAGHAV and pending decisions in ANI v. OpenAI will establish binding precedent establishing whether Indian courts can stretch existing doctrines to accommodate machine-generated and machine-assisted content, or whether legislative amendment proves necessary. The Copyright Act, 1957 requires amendment to address authorship of purely AI-generated works through creation of sui generis rights distinct from traditional copyright, establish statutory frameworks governing licensing of copyrighted works for AI training, and clarify ownership of works generated through human-AI collaboration. The India’s Ministry of Commerce expert panel examining AI copyright implications represents governmental recognition that policy evolution proves essential. Without legislative reform, courts will confront irreconcilable tensions between encouraging technological innovation in AI development and protecting rights of human creators whose works comprise AI training datasets. India’s policy choice in the coming months will establish global precedent regarding whether developing nations embrace expansive AI innovation incentives or prioritize creator protection through stringent copyright enforcement. This choice will substantially influence India’s competitive positioning in global AI markets, the viability of domestic content creation industries, and alignment with international copyright norms under the TRIPS Agreement and WIPO treaties. The convergence of imminent court decisions and pending policy panel recommendations creates an unprecedented opportunity to establish coherent legal framework for AI and intellectual property that protects both innovation and creative rights.