Introduction
On 26 March 2026, the Delhi High Court faced a question that sits right in the middle of modern copyright law, whether a musical composition made entirely by artificial intelligence can actually get copyright protection. In Tarun Chaudhary & Anr. v. Kuldeep Meena & Ors. (April 2026), the plaintiff asked for an injunction, to stop what he claimed was infringement of a song called ‘Teri Yaadon Ki Chadar Odhe’. During the proceedings it came out that the lyrics were written by a human, but the music or the actual musical composition was generated using SUNO AI. Justice Tushar Rao Gedela did not grant interim relief, and he said the Court was not convinced, at least prima facie, that copyright exists in the AI made composition. Since the issue is new, the Court impleaded the Registrar of Copyright as a defendant to figure out whether AI generated works can qualify for protection under Indian law. This case shows that the Indian copyright law rests on the idea that there is a human author, but generative AI keeps producing creative outputs without one.
The Dispute Before the Court
The plaintiffs claimed they owned the song based on an assignment from its original creator, and they wanted a permanent injunction, removal of what they said was infringing content, disclosure of revenues, with damages around ₹2.5 crore. However, the plaint also admitted that the musical composition was made using SUNO AI , and that the ‘human involvement’ was only limited to giving prompts.
In defendant’s view, there was no copyright that can really exist in a composition like this. They also leaned on SUNO AI’s terms of service, pointing out that those terms do not even promise that the output will count as their own work, or that they will be protected in law. According to them, if the original creator possessed no enforceable copyright, a subsequent assignee could not acquire exclusive rights that never legally existed.
The Court made several notable observations. For example, it said that while the lyrics could possibly be treated as a human-made literary work, there was no clear proof of actual human creativity in the musical composition itself. The Court also flagged uncertainty about what exactly the copyright claim is really about whether it is about lyrics, composition, sound recording, or the whole song in general. Most importantly, the Court did not grant an ex-parte injunction , because whether copyright exists in the AI-generated composition still remains doubtful.
The Legal Vacuum in Indian Copyright Law
The dispute shows the limits of the Copyright Act, 1957. Section 2(d)(vi) talks about the author of a computer-generated work as ‘the person who causes the work to be created’. On first look, a user who types prompts into an AI system might seem to tick that box, like they are the one causing it. But when we try to apply this to modern generative AI, then the problems start to stack up.
The first problem is about authorship. Even if users supply prompts, they don’t always have real control over the exact expressive content of what comes out. A prompt can signal a concept, yes, but the AI, it ends up choosing how that concept is actually spelled out. So it becomes hard to say that the user is the true author of the resulting work.
The second problem is originality. In Eastern Book Company v. D.B. Modak (2008), the Supreme Court said that copyright guards the expression of thought where the author’s skill and judgment are involved. Yet purely AI-generated outputs don’t really fit that idea, because the expressive parts come from a machine, not from human intellectual effort.
The third issue is more historical. Section 2(d)(vi) was added when computer-generated works were mostly taken as automated outputs like compilations, reports, or databases. It was not really built with systems in mind that can independently generate music, literature, or visual art. As a result the law ends up giving very thin guidance, on how authorship should be placed in the time of generative AI.
International Approaches
Recent developments abroad show how different jurisdictions deal with similar questions, even if they’re doing it in their own particular way.
In Thaler v. Perlmutter, the United States courts declined copyright registration for an artwork entirely made by an AI system called The Creativity Machine. The courts said copyright law assumes there must be human authorship, and that the works produced using only artificial intelligence do not get protection. In March 2026, the United States Supreme Court refused to hear the appeal ,so the position stays as it was. The U.S. Copyright Office has since reinforced this stance in its detailed Copyright and AI reports, particularly Part 2 (January 2025), which concluded that prompts alone do not confer authorship and that purely AI-generated works remain uncopyrightable.
Worth noting though, the decision didn’t really answer how much human involvement counts as ‘enough’ to establish authorship. The whole case was about a work that was described as being created without any human creative input. So the line between AI-generated and AI-assisted works is still blurry and not settled.
The European Union, via the Court of Justice (CJEU) has built a harmonised setup. In Infopaq (2009), the CJEU set out that a work can be protected as the “author’s own intellectual creation” and then, later on, the same test was expanded across all categories of works. In Painer (2011), the Court explained that originality means the work has to carry the ‘author’s personality’ via ‘free and creative choices’, the sort of thing that leaves a “personal touch”. More recently, Cofemel (2019) said there is no need for any aesthetic or artistic merit requirement, so the only real condition is originality together with objective identifiability. And then, Brompton Bicycle (2020) reaffirmed that even functional forms can be protected, provided the maker had “room for creative freedom” and the design was not imposed just by technical function. This situation is useful for AI thinking, because if an output’s expression is essentially driven by a machine’s probabilistic mechanics, and there is no genuine human space for creative choice, it would probably fail the originality test.
The United Kingdom takes a different route. Section 9(3) of the Copyright, Designs and Patents Act 1988 says that for computer-generated works, the author is ‘the person by whom the arrangements necessary for the creation of the work are undertaken’. While this provision came before generative AI, it still gives a more direct statutory anchor to recognise authorship in computer-generated outputs than what Indian law currently offers.
So overall these contrasting approaches make it clear that AI authorship is not fully settled anywhere. India now faces a contemporary development to which its copyright law must respond.
The Contractual Gap
The case also points to the limits of private contracts. SUNO AI’s Community Guidelines say commercial use of songs made via paid subscriptions is allowed, but they also clearly warn that copyright protection cannot really be assured. The platform admits that copyright offices or authorities might turn down protection for works that are mostly produced with AI.
So there’s this big gap, between what a contract seems to grant you, and what copyright law actually recognizes. In other words, an AI platform can give users rights over the generated files by agreement, but it cannot conjure copyright protection if the law won’t recognize it in the first place. And if courts later decide that AI-generated compositions simply aren’t copyrightable, then users might still have contractual remedies against the platform, while having no real, enforceable intellectual property rights against everyone else.
People may get told they ‘own’ the AI-generated content, yet they can’t really stop other parties from copying or using it. Therefore, the problem is not only who owns the work, but whether any copyright exists from beginning.
Conclusion
The Delhi High Court’s choice to implead the Registrar of Copyright points to a wider reality, the existing copyright law doesn’t give a clear response to the problems that generative AI throws in the mix. At last, it come down to judicial interpretation, some administrative guidance that actually works in practice, or even legislative reform, depending on how the courts want to frame it.
India is at crossroads now. It could lean toward the American route, by only narrowing protection to works that involve meaningful human authorship, or it could lean closer to the United Kingdom’s system for computer-generated works. It may also try to build a whole new solution tailored to its own legal landscape and policy goals.
Until more clarity shows up, disputes like Tarun Chaudhary are likely to keep coming up more often. This case works as a reminder that copyright law was designed around the idea of a human creator. Generative AI really disrupts that premise at the core. And as technology keeps reshaping how creative production happens, Indian law has to decide whether authorship can exist without a human author, and if it cannot, then what happens next to the outputs that AI creates.
Authored by: Mr. Sakkcham Singh Parmaar, Jindal Global Law School, O.P. Jindal Global University
Leave a Reply