Abstract

With the increasing presence of AI in our lives, and its unavoidable ability to replace human thought, it is of much importance that the capability of Indian Laws in this aspect is assessed. Rooted in the assumption of human authorship, the Indian Laws fail to account for AI as an author of a creation.   This blog examines the statutory provisions governing authorship, ownership, and originality, analysing their incompatibility with non-human creators. It also analyses judicial interpretation, particularly the requirement of skill, judgment, and labour, to demonstrate why purely AI-generated works fall outside existing protections. A comparative study of the United Kingdom, United States, and European Union is also given, which shows varying approaches, yet a consistent reluctance to recognise AI as an author. The blog further goes through the theoretical limits of legal personhood and the distinction between AI as a tool and AI as an autonomous creator. It concludes by evaluating three policy alternatives and argues for urgent legislative clarification to prevent legal uncertainty and economic inefficiency.

Elliana Esquivel, an artist, had been making art since she was 16. Much of the work that had been scraped for training different Artificial Intelligence software was created during a period when she was experiencing homelessness, and art was her way out. On a rather normal day, she went to haveibeentrained.com to check if the suspicions she had over her creations being stolen were true. She searched her name and watched the work of her life appear on the screen, without her consent. When she discovered her almost entire portfolio had been taken and used to train the Artificial Intelligence (AI) dataset LAION-5B,[i] she said: “To see that it was scraped, and that people are just reproducing it for free, with no context as to how it was made or why it was made — that was a really sobering moment.”[ii]

As much as we would want it to be, unfortunately, Elliana’s story is not unique. Artists across the world are posed with the same question that now faces the Indian Judicial and Legislative system: Do Copyright laws recognise the artists of AI-generated work? The Copyright Laws in India are mainly specified and covered by the Indian Copyright Act of 1957 (the Act) and its subsequent amendments. This law was structured and laid down around a human author, assuming that there would not be a need for a provision that assigns copyright to AI, or its owner, or the human behind the prompts. This assumption is no longer safe.

Authorship Under the Indian Copyright Act, 1957

Section 2(d) of the Act defines authors of various creations like literary and dramatic works, sound recordings, musical works, artistic works, cinematograph films, etc.[iii] However, all of these authors are presumed to be humans. In works generated by a computer, the ‘person’ who causes the work to be created is considered the author. In Indian law, individuals and legal entities can be classified as persons; however, AI is neither. Section 17 of the Act further states that the author of a work shall be the first owner of copyright.[iv] There is no provision for ownership to be vested in a non-person, a machine, or a process. Section 57 takes this assumption to the next level by granting authors the right to claim authorship and object to the distortion of their work.[v] These rights survive death and are passed to the legal heirs because they are rooted in reputation, dignity, and personal stake. AI can have none of this.

Section 13 of the Act is considerably significant when it comes to the originality of a work; however, the Act nowhere defines originality.[vi] This gap in definition was filled by the Hon’ble Supreme Court in Eastern Book Company v D.B. Modak (2008).[vii] In its judgment, the Court ruled that to claim copyright, a work must include the skill, judgment, and labour of an author.[viii] The Court drew a path away from the conventional sweat-of-the-brow standard and the creative spark test and ruled that originality under Indian law is not an isolated attribute, but a relational aspect that connects the work produced with the mind that produces the work. In the absence of such a mind, the threshold of originality under this Section stands unfulfilled, because AI is not capable of judgment, skills, or labour.

How are Other Jurisdictions Responding?

The Indian Copyright Act leaves AI-generated work in a legal vacuum with no originality, no author, and no ownership. However, this vacuum is not inevitable. India is not alone in facing this problem, but might increasingly be alone in finding a solution.

Section 9(3) of the Copyright, Designs, and Patents Act 1988, which governs the copyright laws of the United Kingdom, already provides a provision for ownership for computer-generated work, designating the person who arranges for the outcome of such work as the owner of the copyright.[ix] This has been judicially interpreted only once, in Nova Productions Ltd v Mazooma Games Ltd, where the developer of a game was provided the copyright over the user.[x] However, this was inferred in the context of video games and not generative AI, leaving this provision under active review and far from total implementation.[xi]

Courts in the United States have consistently refused registration for work that is purely AI-generated, establishing the importance of substantial human involvement.[xii] This stance has been reaffirmed at every level of the judiciary.[xiii] However, the US draws one distinction worth noting. Wherever a human has exercised creative control over a work assisted by AI, the copyright would be granted only for the human contribution, and the AI stands unrecognised.[xiv] This is mainly because the Copyright Office clarified that direction, selection, and arrangement by a human mind is non-negotiable.[xv]

The Court of Justice of the European Union (EU), through recent cases, has emphasised the importance of an ‘intellectual work’ and ‘personal contribution’ of the author in the creative process, excluding purely AI-generated work from copyright.[xvi] The EU AI Act provides for the obligations around AI without touching on the ownership part.[xvii] The state of affairs in the EU is relevant to India because it demonstrates that the regulation of AI and resolving copyright ownership are two different but complementary aspects that should be legislated on separately.[xviii]

Can a Non-Person Be an Author?

Countries around the world have been following the legal personhood theory. Legal personhood essentially allows a person, corporation, or other entity to engage in the legal system. The pertinent question is: if corporations can be designated as a legal person, can AI be? The Hohfeldian framework requires that a rights holder be capable of bearing duties. AI, however, cannot bear duties. It cannot be held liable for infringement, enter into a contract, or face any legal consequence for that matter. Whenever a corporation holds a right, it can be traced to actual humans who would be held accountable. Granting copyright to AI would establish a right with no duty bearer — a monopoly that the machine owner would enjoy.

The involvement of AI in creative production has two principal dimensions. First, where AI functions as a tool — for instance, where a novelist uses autocomplete, or a designer uses an AI filter. Here, most of the work is done by the human, with AI acting as an aiding mechanism. This is instrumental creativity, and existing copyright laws have accommodated this without raising jurisprudential concerns. Second, and problematically, is the autonomous creativity of AI — where AI is not merely a tool, but an originator. When a user enters a prompt and receives a fully composed work without specifying dimensions or making creative choices, the AI is responsible for most of the work and the human simply provides an occasion. As the US Copyright Office framed the question, the key enquiry is whether the work is ‘basically one of human authorship, with the computer merely being an assisting instrument, or whether the traditional elements of authorship… were actually conceived and executed not by man but by a machine.’ [xix] AI has no self to project — it converts input data into numerical representations, processes them through layered mathematical functions, and produces an output by selecting the sequence of values with the highest predicted probability given the training distribution. There is, in short, no personality to present.

The Indian Legislative Gap and Its Consequences

In the present times, with the omnipresence of AI in every profession, an overarching need for clarity on this issue arises. If the current ambiguity continues, it will result in a multiplicity of arguments and interpretations, leading to further confusion. Almost all of the creative work we see today has a monetary motivation behind its creation. With no clarity on this aspect, who profits from AI-generated works? Is it the developer? The tech company? The user, or the public domain? Each of these claimants has a plausible argument. The developer built the system. The tech company owns it. The user provided the creative input. And if none of them qualify under existing law, the work belongs to no one — meaning it belongs to everyone — and therefore no one can monetise it with any legal certainty.

Commercially, the stakes are very high. In the future, music labels may use AI to compose songs, publishing houses to draft manuscripts, and advertising agencies to generate campaigns. The uncertainty itself is a form of market failure. When ownership is unclear, licensing becomes impossible, and enforcement becomes unenforceable. Investment becomes a risk that rational commercial actors will stall before making.

India’s creative economy is growing rapidly, and AI already constitutes a substantial part of it. Without legislative intervention on this issue, the courts will be left to draw answers to questions that Parliament should have already answered.

The Way Forward

With the current status of law and the analysis of provisions in other jurisdictions, India has three broad options when it comes to ownership of AI-generated work.

The first option is to vest ownership in the human who prompts or directs the AI. This causes the least disruption to existing laws and is in line with current provisions. The Act already recognises the person who creates work through a computer as the owner. Integrating this option requires not a complete overhaul, but a broader legislative interpretation. However, this model carries a drawback: it fails to account for instances where human input is minimal, shifting the software from instrumental to autonomous creativity. As the US Copyright Office’s Review Board observed, mere refinement of prompt commands cannot result in copyrightable authorship because ‘the traditional elements of authorship are determined and executed by the technology — not the human user.’[xx] The Board also made clear that a human may claim authorship of visible human-authored modifications to an AI-generated work, but must disclaim the AI-generated portions.[xxi]

The second alternative is to vest copyright in the company that developed the AI software. This recognises the intellectual and commercial effort that developers invest in building AI systems. However, the concern here is that it effectively hands a monopoly to large tech corporations overall output their systems produce, disregarding the user’s contribution and raising significant equity concerns.

A third course of action would be to allow AI-generated work to enter the public domain directly. This is the most consistent approach with authorship theory: if copyright exists to protect the personal expression of a human mind, and no such mind is present, the work logically belongs to no one in particular, and therefore to everyone. The practical consequence, however, is that it removes any commercial incentive to invest in AI-generated creative production. As the legislative history of the UK’s Section 9(3) illustrates, industry participants have long maintained that the absence of copyright protection for non-human outputs poses a risk to investment and disincentivises further technological development.[xxii]

All of the paths discussed above carry their own advantages and disadvantages. Choosing any of them could shape the copyright laws of the country, and should therefore be chosen accordingly.

Conclusion

From Elliana Esquivel watching her life’s work disappear into an AI dataset,[xxiii] to a music label unable to enforce rights over an AI-composed track, the consequences of legislative silence are nearing us as time passes. The Copyright Act of 1957 was built on an assumption of a human mind behind every creative work — an assumption that has gradually ceased to hold. Of the three models available, vesting ownership in the human who prompts or directs the AI offers the most pragmatic interim solution. But it must be interim, not permanent. A Parliamentary amendment to Section 2(d) is not optional but long overdue. Indian courts will face this question sooner than our legislators have anticipated. When that moment arrives, the a

[i]Emma Korynta, ‘Art will never be the same: Charlotte artist has her work scraped by AI’ (WCNC Charlotte, 20 December 2023) <https://www.wcnc.com/article/features/originals/charlotte-artist-elliana-esquivel-artificial-intelligence-ai-scrape-artwork/275-b7c79345-b9cf-4dd4-b685-459515f6c25f> accessed 8 April 2026.

[ii]ibid.

[iii]Indian Copyright Act 1957, s 2(d).

[iv]ibid s 17.

[v]ibid s 57.

[vi]ibid s 13.

[vii]Eastern Book Company & Ors v D.B. Modak & Anr (2008) 1 SCC 1 (SC India).

[viii]ibid. The Court adopted a middle path between the ‘sweat of the brow’ doctrine and the ‘modicum of creativity’ standard, requiring that a work embody the author’s ‘skill, judgment and labour’ with at least minimal creativity.

[ix]Copyright, Designs and Patents Act 1988 (UK), s 9(3): ‘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.’

[x]Nova Productions Ltd v Mazooma Games Ltd [2006] EWHC 24 (Ch). The court held that the software programmer — not the user — was the author under s 9(3) because he devised the game’s visual elements and logic; the player had merely played the game.

[xi]A&O Shearman, ‘Ownership of AI-Generated Content in the UK’ (2025) <https://www.aoshearman.com/en/insights/ownership-of-ai-generated-content-in-the-uk> accessed 8 April 2026.

[xii]US Copyright Office, ‘Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence’ (2023) 88 Fed Reg 16190.

[xiii]US Copyright Office, Copyright and Artificial Intelligence, Part 2: Copyrightability (January 2025) 4.

[xiv]ibid. The Office distinguishes between AI used as a creative tool by a human author (which may be copyrightable) and AI used as a stand-in for human creativity (which is not).

[xv]US Copyright Office (n 12) 16190–16193: ‘Applicants should not list an AI technology or the company that provided it as an author or co-author simply because they used it when creating their work.’

[xvi]Infopaq International A/S v Danske Dagblades Forening (C-5/08) [2009] ECR I-6569 (CJEU).

[xvii]AIPPI, ‘Approaches to IP Protection for Works Generated by Artificial Intelligence: European Standards’ (April 2025) <https://www.aippi.org/news/approaches-to-ip-protection-for-works-generated-by-artificial-intelligence-european-standards/> accessed 8 April 2026.

[xviii]European Parliament Think Tank, ‘Copyright of AI-Generated Works: Approaches in the EU and Beyond’ (EPRS Briefing, December 2025) PE 782.585.

[xix]US Copyright Office (n 12) 16192–16193.

[xx]US Copyright Office (n 13) 4–5.

[xxi]US Copyright Office Review Board, Decision on ‘Théâtre D’Opéra Spatial’ (September 2023); see also Sterne Kessler, ‘The US Copyright Office’s Position on the Copyrightability of Works Made with the Assistance of Generative AI’ (2024) <https://www.sternekessler.com> accessed 8 April 2026.

[xxii]Authors Alliance, ‘The UK’s Curious Case of Copyright for AI-Generated Works: What Section 9(3) Means Today’ (19 May 2025) <https://www.authorsalliance.org/2025/05/19/the-uks-curious-case-of-copyright-for-ai-generated-works-what-section-93-means-today/> accessed 8 April 2026.

[xxiii]Spawning AI, ‘Have I Been Trained?’ <https://haveibeentrained.com> accessed 8 April 2026; see also PetaPixel, ‘You Can Now Check if Your Photos Were Used to Train AI Image Generators’ (19 September 2022).