Algorithmic Artistry - Reconciling Generative AI, Authorship, And Copyright In The Indian Music Industry

Author: LawSphereX

Category: Intellectual Property

Published: 2026-08-23T17:46:47.087789+00:00

Updated: 2026-08-26T17:12:20.564+00:00

ALGORITHMIC ARTISTRY: RECONCILING GENERATIVE

AI, AUTHORSHIP, AND COPYRIGHT IN THE INDIAN

MUSIC INDUSTRY

Chelsy Mehta, National Law Institute University, Bhopal

ABSTRACT

This paper dives into the tough questions generative artificial intelligence(AI) raises for India’s copyright system, particularly in music. It contends that the Copyright Act of 1957, rooted in human notions of ‘author’ and ‘originality’, falls short when dealing with AI-created tunes. By examining key legal clauses, Indian rulings like the RAGHAV mess and the unfolding ANI v OpenAI battle, plus lessons from abroad, it unravels the chaos around ownership, copying, and accountability. It also weighs the bigger picture, social, financial, and philosophical, for the artists. Wrapping up, it pitches a comprehensive fix for India: precise law tweaks, a courtstandard ‘Significant Human Input’ check, and solid, collaborative setups for licenses and disclosure to nurture both tech progress and artistic spirit.

Keywords: Generative AI, Copyright Law, AI-Generated Content, Ownership Rights, Authorship, Originality, Infringement, Human Input Threshold, Licensing Frameworks, Technological Innovation, and Artistic Rights.

I.

Introduction: The New Cadence of Creation Music-making is facing a radical overhaul, thanks to the explosive growth of generative AI. Apps like Suno, Udio, and AIVA are putting composition in everyone’s hands, turning basic text descriptions into polished songs with everything from melodies to vocals. These are not just helpers like editing software or synthesisers, they are true creative allies. They pore over massive song libraries to grasp nuances in melody lines, chord progressions, rhythms, and vibes, then craft originals that either nod to classics or venture into uncharted territory. The upside? It speeds things up, welcomes novices, and sparks inspiration when ideas run dry. With how user-friendly these tools are, we are witnessing an inundation of new music unlike anything before. This is not hype- AI tracks are already streaming on platforms like Spotify.

But this innovative beat is hitting sour notes legally. At its core, the tech lets machines handle the artistic heavy lifting humans once owned, upending the rudiments of intellectual property. As adoption skyrockets, a wave of ‘orphan works’, which basically are copyrighted materials for which the owner cannot be identified or located after a diligent search, creates a legal dilemma for potential users who cannot obtain permission to use the work without infringing copyright, which can be legally fuzzy, overwhelming a copyright framework never built for this scale. It’s no ivory-tower issue; it’s sparking real chaos in disputes over payments, rights, and violations that our laws are not geared to sort out. The core clash stems from India’s Copyright Act1, crafted in an analogue age centred on people.

Its definitions of ‘author’ and ‘originality’ hinge on a human applying talent, effort, and discernment. Generative AI, powered by intricate code and learning layers, shatters that mould. This piece tackles the thorny problems this lacuna creates: Who owns an AI-spun melody? Is training on protected songs a theft? How to assign fault if the result echoes something copyrighted? Above all, how should India evolve its rules to embrace this tool without undermining the human ingenuity it’s meant to shield? The hazy status quo on AI music is unsustainable, breeding doubt that deters tech investment and endangers artists’ rights and earnings. India must forge ahead with smart updates to its copyright setup. Ignoring it could mean unchecked tech dominance or stifled growth via rigid old interpretations. I shall map this out by first laying the groundwork of Indian copyright–

1 Copyright Act 1957.

originality and authorship essentials. Then, I’ll dissect the AI authorship enigma, parsing the Act’s language and the ambiguous RAGHAV precedent2. Next comes the infringement duo: data inputs for training and outputs that mimic. I’ll highlight the crucial ANI v OpenAI case3. A global scan of US, UK, and EU strategies offers reform ideas. I’ll also reflect on the deeper

ethical, philosophical, and economic ripples in music4 5. Finally, I’ll outline a rounded set of proposals for India’s laws, courts, and policies to craft a system suited to algorithm-driven art. II.

The Pillars of Protection: Originality and Authorship in Indian Copyright Law The Act protects defined creative forms. Section 2(y)6 includes literary, dramatic, musical, artistic items, movies, and audio. Here, ‘musical work’ matters most- Section 2(p)7 calls it music plus any notation, excluding lyrics or performances meant to accompany it. That’s vital: a song’s protection splits the composition (tune, harmony, beat) from lyrics(as literature) and recordings(separately). AI often creates fresh compositions outright, striking as musical work safeguards. Copyright demands ‘originality’ per Section 13(1)(a)8, but without a definition, courts decide accordingly. Long influenced by UK views, India once used ‘sweat of the brow’- protection based on hard work, regardless of flair. A directory might qualify for the grunt work alone.

That was flipped in the Supreme Court’s Eastern Book Company v D. B. Modak9 decision, scrapping ‘sweat’ for favouring toil over true invention. Now, it’s ‘skill and judgement’ or a ‘modicum of creativity’: the work must stem from the author, not a copy, with input that’s not trivial or mechanical. For AI music, this pivot is ironic. The ruling that updated the law to prize creativity now hurdles AI claims. ‘Sweat’ might have covered AI’s computational grind and costs. But the new test requires human spark- missing in pure algorithm output. Thus, modern jurisprudence 2 The matter of Ankit Sahni’s copyright registration for the AI artwork ‘Suryast’ (created by an AI named RAGHAV). The application was initially rejected, then granted with Sahni as a co-author, and subsequently reviewed.

3 ANI Media (P) Ltd v Open AI Inc 2024 SCC OnLine Del 8120

4 'Addressing ethical considerations in AI development by Indian companies' (Mold Stud) accessed 6 October 5 'AI in Music Creation and the Ethical Challenges of It' (Medium, 15 November 2023) accessed 6 October 2025

6 Copyright Act 1957, s 2(y)

7 Copyright Act 1957, s 2(p)

8 Copyright Act 1957, s 13(1)(a)

9 Eastern Book Company v DB Modak (2008) 1 SCC 1

unwittingly bars machine-made art form protection. Authorship is decidedly human-oriented. Section 2(d)10 ties it to roles like ‘composer’ for music, assuming personal creation—section 1711 vests initial ownership in the author (with exceptions), envisioning a living individual. When AI composes solo, the system falters, birthing works without a legal originator. III.

The Algorithmic Composer: Deconstructing Authorship in AI-Generated Music AI’s authorship riddle is central: no author, no copyright, potentially public domain status that chills innovation. The Act’s Section 2(d)(vi) seems to help, but mostly muddies the waters. Interpreting Section 2(d)(vi): Who is the ‘Person Who Causes the Work to be Created’? Added in ‘94 for computer outputs, it deems the author the one causing the creation. Forward thinking, but for advanced AI, it’s vague, yielding three interpretations. i. The Developer as Author Perhaps builders/trainers cause it, arranging essentials as in UK law12. Their work enables output. Yet, they lack control over specific results- users prompt those, so authorship does not stick easily. ii. The User as Author More persuasively, users direct via prompts, iterations, and selections. They spark intent.

But the US Copyright Office counters that prompts are ideas, not expression, AI chooses notes, etc. , so users might not qualify. iii. A Human-AI Duet: Exploring Joint Authorship Or we can view it as a collaboration. Section 2(z)13 defines joint works as blended contributions. Fits the dynamic, but AI’s non-person status blocks it, as seen in RAGHAV.

10 Copyright Act 1957, s 2(d)

11 Copyright Act 1957, s 17

12 Copyright, Designs and Patents Act 1988

13 Copyright Act 1957, s 2(z)

RAGHAV highlighted this confusion. In 2020, Ankit Sahni applied for ‘Suryast’, an AI artwork, first crediting RAGHAV alone, but was denied, as machines are not authors. A redo with co-authorship got initial approval, a global first. But quick reversal via notice underscored the framework’s inadequacy, leaving authorship in flux14. This hole means AI music might get zero protection, going public instantly. Good for access, bad for motivating AI developmentif outputs cannot monetise, why invest? It also encourages gaming: exaggerating human roles for registration, overwhelming offices with dubious claims and sparking inconsistencies and suits. AI challenges enforcement twice: training inputs and similar outputs. Models train on copied copyrighted songs, seemingly violating Section 1415’s reproduction rights.

This fuels ANI v OpenAI, where news and music groups (via IMI) allege unauthorised use. It’s a showdown of models: licensing-based creatives vs data-gobbling tech. A plaintiff win mandates deals, redirecting funds, whereas a defendant win empowers extraction, devaulting content. Courts here shape policy. The US's ‘fair use’ is broad; India’s ‘fair dealing’ lists specifics like research or review. It outputs a match via ‘unmistakable impression’ (R. G. Anand16), its infringement. Blame could hit users (prompts), developers (tools/users), and platforms (hosting). IT Section 7917 offers safe harbours with diligence, but evolving rulings demand more, keeping liability knotty. IV.

Global Lessons and Ethical Ripples India can learn from others’ varied tactics. The US insists on human creation; Thaler v Perlmutter18 confirmed that the works generated from AI alone won’t qualify. But humantweaked AI outputs can, disclaiming machine parts. Now, what about prompts alone? Certainly, not. The UK’s 1988 Act assigns authorship to the arranger for computer works, prioritising ownership to fuel investment- practical over purist. The EU emphasises rules like the AI Act’s data summaries for enforcement19 20, plus the EU’s Directive human-creation standard21 provides for accountability via governance. These approaches vary: In India, AI

14 'Authorship and Ownership in Copyright Law' (The Legal School) accessed 6 October 2025

15 Copyright Act 1957, s 14

16 R G Anand v Delux Films (1978) 4 SCC 118

17 Information Technology Act 2000, s 79

18 Thaler v Perlmutter 2023 WL 5331132 (DDC 2023)

19 'EU AI Act: first regulation on artificial intelligence' (European Parliament, 19 February 2025) accessed 6 October 2025 20 'EU AI Act: How Far Will EU Copyright Principles Extend?' (Morgan Lewis, February 2024) accessed 6 October 2025 21 Directive (EU) 2019/790 of the European Parliament and of the Council of 17 April 2019 on copyright and related rights in the Digital Single Market OJ L130/92

can’t be author implicitly, with ambiguous ‘person who causes’ per Copyright Act Section 2(d)(vi), is human-creation but vague. While in the US, it is not explicitly inferred as it is not copyrightable without human input, via case laws and guidance, and is tied to human incentive. On the other hand, in the UK, the law explicitly handles “computer-generated” works. It treats the author as the person who arranged for the work’s creation. In practice, it means that the programmer or user behind the AI is considered the author. Finally, EU rules focus on originality from a human author. Courts say a work must reflect the author’s own intellectual creation, so purely AI-generated content (with no real human creative input) generally does not qualify for copyright. It can be argued that each jurisdiction centres on the human element at the core.

AI itself is not the author, and significant human creativity or direction is needed before a work can get copyright protection in India, the US, the UK, or the EU. These are not dry debates; rather, these impact incomes, ethics, and art’s essence. Musicians fear displacement in stock music as AI’s speed undercuts markets. Projections show a 25% income loss by 2028. In addition to this, the content overload buries humans, shrinks royalties, while commoditising music. Music thrives on human emotion, as AI lacks this segment, yielding hollow averages. Yet, some see it as a tool for boundary-pushing, like Grimes’ fan collabs. The key ethics associated here are unconsented training exploits that demand transparency and its attribution, and curb deepfakes (as projected in the Arijit Singh ruling on voice rights22).

This traction elongates beyond music to visual arts, as exemplified by the recent Studio Ghibli controversy, where OpenAI’s ChatGPT image generator permitted users to swamp social media with “Ghiblified” images imitating the studio’s hand-drawn anime style pictures, prodding retaliation over potential training on copyrighted Ghibli works without permission and kindering Hayao Miyazaki’s longstanding vilification of AI imagination as “an insult to life itself”. While styles themselves are not copyrightable, the discourse accentuates fears of unauthorised data use eroding artistic integrity, much like musicians worry about AI mimicking protected compositions, stoking up calls for stronger safeguards against stylistic appropriation across creative fields. 22 An unreported ruling concerning the protection of singer Arijit Singh’s voice rights against misuse by AI.

V.

Conclusion and Recommendations Generative artificial intelligence is shaking up the world of creativity in ways we could not have imagined, and it is putting a spotlight on how India’s Copyright Act of 1957 just isn’t keeping pace, a law that was written with the firm belief that only people could create and own intellectual property. When AI steps in to compose music, the Act runs into real trouble trying to sort out who the owner is, what makes something original, and how to deal with possible copying, as we have seen in the messy back-and-forth of the RAGHAV case and the heated courtroom clashes in ANI against OpenAI23. This dilemma isn’t just something for scholars to ponder; it is leading to actual problems that slow down new tech and leave artists wondering if their work is truly protected. We cannot afford to just hang back and hope it all works out.

These issues are showing up everywhere, from corporate strategy sessions to court hearings and even in the day-to-day lives of musicians, which means India really needs to roll up its sleeves and put together a smart, all-around plan. Starting with the laws, updating Section 2(d)24 feels like a must, to separate AI-helped creations, where humans are calling the shots on the big artistic choices and should get full copyright, from those that are mostly the AI’s doing, with barely any human touch. For the stuff that is all AI, maybe bring in a custom-made sui generis right25 that offers some basic protections, like a shorter 25-year term, going to whoever started the whole thing. That keeps human efforts at the top of the heap value-wise, but still lets AI substances find their place in the economy.

On top of that, fortifying Section 5226 with a clear allowance for digging through data to train AI would help, as long as creators get a strong way to say no and keep control. Courts are not off the hook either; they could step in with something like a ‘Significant Human Input’ test to really look at whether someone’s role made a meaningful difference in the final product, digging deeper than just the initial prompt to check for genuine creative effort. Then, on the policy front, we would do well to push for more openness, like making AI companies share 23 'Indian Music Industry files intervention application in ANI v. OpenAI copyright case' (Trademark Lawyer Magazine, 20 March 2025) accessed 6 October 2025

24 Copyright Act 1957, s 2(d)

25 'Unravelling The Enigma: The Quest For Sui Generis Rights In AI Creation' (Law School Policy Review, 27 March 2024) accessed 6 October 2025.

26 Copyright Act 1957, s 52

details on what data they trained on, and set up simple licensing deals through trusted groups such as the Indian Performing Rights Society (IPRS). And let’s not forget about ethics, coming up with rules to handle deepfakes, insisting on labelling AI-made content, and protecting privacy is key. If we weave all this together, India stands a good chance of finding that middle ground where tech breakthroughs and human imagination not only get along but actually lift each other up, creating a cultural scene that’s broader, more welcoming, and fairer across the board.