From Silence to Statutory Clarity: A Comparative Analysis of Copyright Ownership in AI-Generated Works in India & UK

Introduction

In October 2018, an AI-generated portrait Edmond de Belamy sold at Christie’s for USD 432,500. In 2020, the Indian Copyright Office officially registered the artwork Suryast with the AI system RAGHAV which stands for Robust Artificially Intelligent Graphics and Art Visualizer listed as a co-author, the Copyright Office had previously rejected an application that listed the AI as the sole author. They approved it only when a human was added as a joint creator. After the historical approval made global headlines, the Indian Copyright Office backpedaled and issued a withdrawal notice questioning RAGHAV’s legal status as an author. Sahni attempted to registered the same co-authored artwork in other countries. The U.S. Copyright Office firmly rejected it, stating the image lacked the necessary human expression. These events reveal a major global problem: copyright law was originally designed to protect human creativity, but now it is struggling to deal with content created by artificial intelligence (AI).

India and the United Kingdom offer a compelling comparative study. Both inherited the common law tradition, both have statutory provisions that address computer-generated works, and yet their legal ecosystems have produced strikingly different outcomes. The UK has a specific legislative mechanism, Section 9(3) of the Copyright, Design and Patent Act,1988[1] that assigns authorship where no human author exists. India has a parallel provision Section 2(d)(vi) of the Copyright Act 1957 [2]but the surrounding jurisprudence, the originality standard, and the institutional posture have produced something closer to statutory silence than to clarity. This article argues that neither jurisdiction has achieved a coherent framework and that both must confront the same three structural problems: the authorship deficit, the originality paradox, and the ownership vacuum.

The UK Framework: Section 9(3) CDPA 1988—A Legal Fiction Under Strain

Legislative Genesis

Section 9(3) of the Copyright, Designs and Patents Act 1988 [3]provides:

“In the case of a literary, dramatic, musical or artistic work which is computer-generated, the author shall be taken to be the person by whom the arrangements necessary for the creation of the work are undertaken.”

“Computer-generated” is defined in Section 178 of the Act as referring to works “generated by computer in circumstances such that there is no human author of the work”. This provision/law was enacted in the late 1980s as a result of lobbying by the software and computing industries. At that time, software companies feared that if computer-generated output did not receive copyright protection, competitors would freely copy the outputs produced by their computers. This could have dampened corporate investment and interest in AI technology. Consequently, this provision granting copyright protection was introduced. When this law was enacted, it was considered highly significant because it was one of the world’s first copyright laws to specifically address the emergence of Artificial Intelligence.

The statute constructs a causal link between the work and the person responsible for its creation to fill the authorial void. This attribution functions less as recognition of personal creative expression and more as a policy mechanism designed to support investment. The quid pro quo is clear: the “author” under Section 9(3) receives no moral rights (Section 81 CDPA), and the copyright term is reduced from the standard life-plus-seventy years to just fifty years from the end of the calendar year of creation (Section 12(7)).

Judicial Interpretation: Sparse and Reluctant

Despite being on the statute books for nearly four decades, Section 9(3) has received remarkably little judicial interpretation. Only two cases merit discussion.

In Nova Productions Ltd v Mazooma Games Ltd [2007] EWHC 222 (Pat)[4], the High Court held that images generated during gameplay were protectable computer-generated works, and that the programmer not the player, was the author under Section 9(3), because he had written the code and designed the rules governing image production. The player, though instrumental in causing the images to appear, was not considered to have made the “arrangements necessary.”

In THJ Systems Ltd v Sheridan [2005] EWCA Civ 24[5], both parties advanced claims of authorship under Section 9(3) in pleadings concerning financial charts generated by software. However, the Court of Appeal did not interpret Section 9(3) at all. Instead, it focused exclusively on the post-Infopaq originality standard—”free and creative choices” and a “personal touch”—and concluded that the software developer had made sufficient creative choices (selection of colours, font styles, visual design) to qualify as the human author under conventional copyright principles. Once traditional human authorship was found, there was no need to resort to the legal fiction.

These decisions reveal a judicial tendency to locate traditional human authorship wherever possible. Section 9(3) operates as a fallback, not a primary mechanism. Courts begin by asking whether a human can satisfy the originality requirement under Section 1(1)(a), and only turn to Section 9(3) if no such author can be identified. This approach demonstrates a deep reluctance to engage with the provision’s interpretive uncertainties.

The Originality Paradox

The fundamental issue with Section 9(3) is its conflict with the established originality standard. Under the post-Infopaq jurisprudential framework, copyright protection requires a work to reflect the “author’s own intellectual creation”—manifested through human personality, free creative choices, and a personal touch. Because computer-generated works by definition lack a human author, CDPA’s statutory presumption of originality creates an inescapable legal contradiction.

The 2026 Government Review: Towards Repeal?

The UK Government’s Report on Copyright and Artificial Intelligence[6], published in March 2026 under the Data (Use and Access) Act 2025[7], notes that most consultation respondents supported repealing Section 9(3) to remove copyright protection for works created entirely by AI without a human author. While the government agrees that copyright should focus on protecting human creativity rather than wholly AI-generated outputs, it concluded that immediate legislative changes require further evidence on economic impacts. Consequently, Section 9(3) remains in a state of legal uncertainty—currently too unreliable for long-term IP strategies, yet retained for now until additional data is gathered.

The Indian Framework: Section 2(d)(vi)—Statutory Text Without Doctrinal Life

Legislative History and Text

India’s Copyright Act 1957 was amended in 1994 (via the Copyright (Amendment) Act 1994) to introduce Section 2(d)(vi), [8]which provides:

“In relation to any literary, dramatic, musical or artistic work which is computer-generated, the  author means the person who causes the work to be created.”

The language closely parallels the UK’s Section 9(3)—both assign authorship to the person responsible for the arrangements or causation of the work’s creation. However, the Indian provision does not carry the same qualifying language (“in circumstances such that there is no human author of the work”) nor the accompanying limitations on term and moral rights that temper the UK provision.

The RAGHAV Anomaly: Co-Authorship for AI

In November 2020, the Indian Copyright Office registered an artwork titled Suryast—a photograph of a sunset recreated in the style of Van Gogh’s Starry Night by an AI tool called RAGHAV (Robust Artificially Intelligent Graphics and Art Visualizer)—listing both the human creator Ankit Sahni and the AI system RAGHAV as co-authors. [9] The registration was granted despite the Copyright Act containing no provision for non-human co-authors. The US Copyright Office subsequently refused registration for the same work, holding that “the new aspects of the derivative work were generated by RAGHAV rather than Sahni or any other human author”[10]. In India, while the registration status still appears as “registered” on the copyright office website, the matter has not been adjudicated by any Indian court, leaving its legal validity in a zone of uncertainty.

This episode reveals a fundamental institutional disconnect: the Copyright Office appears to have acted without statutory warrant, and no judicial body has yet corrected the anomaly.

India’s Commerce Ministry has declared the current IP regime “well-equipped to protect AI-generated works” and indicated no plan to create a separate category of rights (IJLLR, 2026), a position that sits uneasily with the absence of any jurisprudential foundation for AI authorship.

Originality: The “Skill and Judgement” Standard

The Indian Supreme Court established the originality standard in Eastern Book Company v D.B. Modak (2008) 36 PTC 1 (SC)[11], rejecting both the low “sweat of the brow” threshold and the high American “modicum of creativity” test, and settling on the Canadian standard of “skill and judgement.” Under this test, a work must be a product of the author’s skill and judgement, and the exercise must not be so trivial as to be purely mechanical.

In Navigators Logistics Ltd. v Kashif Qureshi & Ors (2018)[12], a copyright claim on a list compiled by a computer was rejected due to lack of human intervention. The court held that human involvement in the creation process is essential for the grant of copyright protection. This precedent is significant: if an AI system independently generates content with minimal human input—say, a single-line prompt—the Indian courts are unlikely to find the requisite skill and judgement.

The Human-Authorship Imperative

Indian copyright law does not recognise AI as a legal person. Section 17 establishes that the author of a work shall be the first owner of copyright, and the entire statutory scheme—from registration forms requiring the claimant’s name and nationality (Form-XIV) to the scheme of exclusive rights, is built around human or legally recognised persons. In R.G. Anand v Delux Films,[13] the Supreme Court clarified that copyright protects the expression of ideas, not ideas themselves. Applied to AI, this suggests that if a human user contributes original expression through the prompt or an iterative process, they may hold copyright in the resulting work—but the AI itself cannot.

The current position can be summarized: (i) AI itself cannot hold copyright; (ii) original data sources generally lack a claim unless there is direct copying; and (iii) the prompt-giver may hold copyright if their input reflects sufficient creativity and skill (Khurana & Khurana, 2025). This is a framework of exclusion, not one of positive entitlement.

Comparative Analysis: Convergence and Divergence

Statutory Architecture

The UK and India share a statutory mechanism that assigns authorship to the person who causes or arranges the creation of computer-generated work. But the UK provision is a self-contained legal fiction with defined boundaries (reduced term, no moral rights, a trigger condition of no human author). India’s provision is embedded in a definitional clause without those guardrails, making it both broader and more vulnerable to misuse, as the RAGHAV registration demonstrated.

The rights regime further underscores the consequences of these divergent designs. In the UK, moral rights do not attach to computer-generated works, the paternity right under Section 79(2)(c) and the integrity right under Section 81(2)(c) of the CDPA 1988 [14]are expressly excluded. This means the deemed author, typically the programmer, holds only economic rights. India contains no analogous exclusion, and since Section 57 of the Copyright Act confers moral rights upon every “author”, the statutory person who merely causes the work to be created acquires perpetual rights of attribution and integrity despite having no personal creative connection to the work. The term of protection reveals a similar philosophical gap: the UK deliberately limited protection to fifty years, recognising that machine-generated expression does not flow from human personality and thus merits a shorter monopoly. India, however, grants the standard sixty-year term without distinction, treating mechanical output on par with deeply personal human creativity, a generosity that lacks doctrinal justification.

The Originality Problem: A Shared Paradox

Both jurisdictions face the same paradox: their originality standards require human intellectual effort, yet their computer-generated-work provisions contemplate the absence of human authorship. The UK’s “author’s own intellectual creation” test demands a “personal touch” that a machine cannot provide. India’s “skill and judgement” test similarly requires the exercise of human faculties. In both cases, the statutory mechanism for computer-generated works sits in tension with the doctrinal foundation of originality.

The UK has at least acknowledged this tension, the 2026 Government Report explicitly questions whether Section 9(3) remains fit for purpose. India has not. The Commerce Ministry’s assertion that existing law is sufficient is contradicted by the absence of any judicial precedent applying Section 2(d)(vi) to AI-generated works, and by the fact that the originality standard was developed for human, not machine, creativity.

Ownership: The Gap Between Text and Reality

Under UK law, Section 9(3) designates the person making the arrangements as author and first owner. The Nova Productions decision identifies the programmer, not the user, as this person. Under Indian law, Section 2(d)(vi) designates “the person who causes the work to be created,” which could potentially be interpreted as either the programmer/developer or the user/prompter—there is no judicial authority to resolve the ambiguity. Scholars have argued that it need not be only the user; the developer of the AI could also be the person who “causes” the work.  This unresolved ambiguity means that both the AI developer and the user could claim authorship, or neither could establish it to the requisite legal standard.

Proposed Reforms

For India: From Silence to Statutory Clarity

Legislative Amendment of Section 2(d)(vi): The provision should be revised to explicitly address AI-generated works, specifying that the “person who causes the work to be created” must be a natural person or legal entity who has made a “significant human input” into the creative process. This test, proposed by scholars at NLSIU (NLS Forum, 2024), evaluates whether the work would exist in the same tangible form without human intervention. A single-line prompt should not suffice; iterative prompting, curation, selection, and creative direction should.

Explicit Exclusion of AI as Author: Parliament should amend the Act to clarify that AI systems cannot be authors or co-authors, rectifying the RAGHAV anomaly and preventing future registrations of the same kind.

Reduced Term and Conditional Moral Rights: Following the UK model, AI-generated works that qualify for protection should receive a reduced copyright term (e.g., 25–50 years) and the human author should retain moral rights only in proportion to their creative contribution.

Judicial Development of the Originality Standard: Indian courts should articulate a tiered originality standard for AI-involved works: (a) works with substantial human creative input receive full protection under the existing “skill and judgement” test; (b) works with minimal human input (e.g., simple prompts) receive either no protection or a sui generis short-term right; (c) wholly autonomous AI outputs receive no copyright protection.

For the UK: From Legal Fiction to Functional Framework

Clarify or Replace Section 9(3): The Government’s 2026 Report [15]should be followed by legislation that either (a) clarifies Section 9(3) with explicit guidance on what constitutes “arrangements necessary” in the context of generative AI, distinguishing between the AI developer and the user, or (b) replaces Section 9(3) with a new provision that directly addresses AI-generated outputs, calibrating protection to the degree of human involvement.

Resolve the Originality Tension: Any reformed provision must either (a) create a separate, lower originality threshold for computer-generated works (as the current de facto presumption does, but without doctrinal justification), or (b) explicitly require that even computer-generated works must reflect some human intellectual creation, with Section 9(3) serving only where residual human creative choices are present but insufficient for traditional authorship.

Distinguish AI-Assisted from AI-Generated: The distinction between AI-assisted works (where a human uses AI as a creative tool and meets the standard originality threshold) and AI-generated works (where no human author is identifiable) should be codified. This aligns with the emerging judicial practice of treating Section 9(3) as a fallback provision.

Conclusion

The journey from silence to statutory clarity is unfinished in both India and the United Kingdom. The UK was first to legislate, but its Section 9(3) has become a legal fiction in both senses—statutory and rhetorical. India followed with mirrored language but without the doctrinal architecture to give it effect. Both jurisdictions now face the same fundamental challenge: copyright’s human-centric foundations are being tested by machines that can produce culturally and commercially significant works with diminishing human involvement.

The UK is closer to reform, with an active government review and a willingness to confront the inadequacy of Section 9(3). India remains in a posture of declared sufficiency—asserting that existing law is adequate while the RAGHAV registration stands as an unresolved contradiction and the courts have yet to apply Section 2(d)(vi) to AI outputs in any meaningful way.

Neither silence nor a legal fiction is a sustainable position. Statutory clarity requires legislatures to do what courts and administrative offices cannot: define the boundaries of authorship, calibrate originality to the realities of AI-assisted creation, and assign ownership in a manner that incentivises innovation while protecting human creativity. The alternative is a growing zone of legal uncertainty in which rights are claimed but cannot be enforced, works are created but cannot be owned, and the law falls behind the technology it was designed to govern.

Reference(S):

Cases

         Eastern Book Company v D.B. Modak (2008) 36 PTC 1 (SC)

         Navigators Logistics Ltd. v Kashif Qureshi & Ors (2018)

         Nova Productions Ltd v Mazooma Games Ltd [2007] EWHC 222 (Pat)

         R.G. Anand v Delux Films AIR 1978 SC 1613    

         THJ Systems Ltd v Sheridan [2005] EWCA Civ 24

Legislation

        Copyright Act 1957

         Copyright, Design and Patent Act 1988

         Data (Use and Access) Act 2025

Secondary Sources

         Indian Copyright Office, Registration No ROC A-135081/2020

         Intellectual Property Office, Report on Copyright and Artificial Intelligence 2026

         UK Government’s Report on Copyright and Artificial Intelligence

         US Copyright Office Review Board, Decision Letter on SURYAST 11 December 2023

[1] Copyright, Design and Patent Act, 1988

[2] Copyright Act, 1957

[3] Copyright, Design and Patent Act, 1988

[4] Nova Productions Ltd v Mazooma Games Ltd [2007] EWHC 222 (Pat)

[5] THJ Systems Ltd v Sheridan [2005] EWCA Civ 24

[6] UK Government’s Report on Copyright and Artificial Intelligence

[7] Data (Use and Access) Act 2025

[8] Copyright Act, 1957

[9] Indian Copyright Office, Registration No ROC A- 135081/2020

[10] US Copyright Office Review Board, Decision Letter on SURYAST (11 December 2023)

[11] Eastern Book Company v D.B. Modak (2008) 36 PTC 1 (SC)

[12] Navigator Logistic Ltd v Kashif Qureshi (2018)

[13] R.G. Anand v Delux Films AIR 1978 SC 1613

[14] Copyright, Design and Patent Act, 1988

[15] Intellectual Property Office, Report on Copyright and Artificial Intelligence (2026)