Artificial Intelligence And The Future Of Music Copyright - Analysing Fair Use In The Age Of Generative AI

Author: LawSphereX

Category: Technology, Cyber & Privacy

Published: 2026-08-23T18:15:46.044526+00:00

Updated: 2026-08-26T17:12:02.723+00:00

ARTIFICIAL INTELLIGENCE AND THE FUTURE OF

MUSIC COPYRIGHT: ANALYSING FAIR USE IN THE AGE

OF GENERATIVE AI

Payal Wagh, Government Law College, Mumbai

ABSTRACT

Music AI has come a long way. What began as simple audio editing software has grown into systems that compose complete songs in moments, these systems challenge fundamental assumptions about authorship, originality, and fair use that have governed intellectual property for centuries. This paper examines the critical legal tensions arising from AI's use of copyrighted material in training datasets, with particular focus on the landmark litigation between Universal Music Group and Suno Inc. Through comprehensive analysis of U.S. copyright doctrine—specifically sections

106 and 107—this study evaluates whether AI music generation constitutes

fair use or represents infringement that could devastate the creative economy. The research explores four pivotal questions: the scope of reproduction rights violations during AI training, the boundaries of derivative work protection for AI-generated outputs, the evolving definition of "transformative use" in the digital age, and the calculation of market harm when AI systems compete directly with human creators. Drawing from recent U. S. Copyright Office guidance and federal court decisions, this analysis reveals that current legal frameworks and address AI's unique capabilities to create perfect digital copies and generate competing content at Novel scale. This analysis reveals that AI's transformative creative capabilities exist within a complex legal framework where traditional copyright protections may not adequately address the unique challenges posed by algorithmic creativity.

The tension between protecting established creative rights and fostering technological innovation raises fundamental questions about how legal systems should adapt to accommodate both human creators and AI developers in our rapidly evolving digital creative ecosystem. Keywords: Artificial Intelligence, USA Copyright Laws, Fair Use, Reproduction Right, Derivative Work, Transformative Rights, Market Harm.

1. Introduction: AI's Disruption of the Music Industry

The global music industry, valued at over $25 billion annually, faces Hitherto unseen challenges from artificial intelligence technologies that threaten to fundamentally restructure the production, distribution, and monetization of creative works. 1 Music is a copyright-driven business. Revenue comes from licensing songs, streaming platforms, and live performances. AI poses a challenge because it produces music so fast and in such large volumes that it could undercut these established revenue streams. 2 AI's rapid evolution from basic audio editing tools to autonomous composition systems has created immediate legal urgency. Record labels representing major artists like Chuck Berry and Mariah Carey discovered that AI platforms could replicate their artists' musical styles within seconds. This raised fears about market saturation with AI-generated competitors.

3 The labels responded by going to court. UMG, Sony Music, and Warner Records filed lawsuits against AI companies like Suno, claiming unauthorized use of copyrighted material in AI training. 4 This creates serious financial problems. AI threatens multiple revenue sources: streaming, sales, and licensing deals that keep the music industry running. 5The core issue is substitution - if AI music can replace human-created songs, it devalues original work while raising copyright concerns. This forces courts to tackle complex questions about digital fair use and market damage. The UMG v. Suno case highlights the central dispute The lawsuit between UMG and Suno captures the heart of this debate. The big question is whether AI companies can legally use copyrighted songs to train their systems, or if this amounts to stealing that could ruin the music industry.

2. AI Music Generation and Copyright Implications

2.1. How AI Systems Create Legal Liability

1 International Federation of the Phonographic Industry, Global Music Report 2024: State of the Industry 8-12 (2024).

2 Id. at 2-5.

3 Complaint at 3-4, UMG Recordings, Inc. v. Suno, Inc., No. 1:24-cv-11611 (D. Mass. June 24, 2024).

4 Id. at 1-5.

5 Oleksandr Bulayenko et al., AI Music Outputs: Challenges to the Copyright Legal Framework, reCreating Europe Report 21-25 (2022).

Modern AI music generation systems like Suno operate through the copying, that creates multiple points of copyright liability. 6 The training process directly implicates the reproduction right under Section 106(1) through sequential steps: (1) downloading and storing,7 (2) preprocessing and converting,8 (3) maintaining. 9 This copying clearly satisfies the reproduction element under established precedent. In Perfect 10, Inc. v. Amazon. com, Inc. , the Ninth Circuit held that storing copyrighted images on servers constitutes reproduction regardless of ultimate purpose. 10 Similarly, MAI Systems Corp. v. Peak Computer, Inc. established that loading copyrighted material into computer memory creates actionable reproduction. 11 AI training requires both persistent storage and memorybased processing of copyrighted works, triggering reproduction rights violations at multiple stages.

Commercial AI platforms like Suno monetize this copying through subscription services, directly profiting from reproduction of copyrighted works. 12 This commercial exploitation eliminates potential fair use defenses available for non-commercial copying and creates immediate economic harm to rights holders.

2.2. Why AI Copying Differs from Human Creativity

The U.S. Copyright Office explicitly distinguishes AI processing from human learning: "AI learning is different from human learning in ways that are material to the copyright analysis" because generative AI training involves the creation of perfect copies with the ability to analyse works nearly instantaneously."13 This technological capability transforms influence

6 UMG Recordings, Inc. v. Suno, Inc., supra note 3, at 3-4 (D. Mass. June 24, 2024).

7 Id. at 15.

8 U.S. COPYRIGHT OFFICE, COPYRIGHT AND ARTIFICIAL INTELLIGENCE, PART 3: GENERATIVE

AI TRAINING, at 25-26 (pre-publication version, May 2025).

9 Id. at 27.

10 Perfect 10, Inc. v. Amazon.com, Inc., 508 F.3d 1146, 1160-61 (9th Cir. 2007).

11 MAI Sys. Corp. v. Peak Comput., Inc., 991 F.2d 511, 518 (9th Cir. 1993).

12 UMG Recordings, Inc. v. Suno, Inc., supra note 60, at 21.

13 U.S. COPYRIGHT OFFICE, supra note 62, at 48.

into potential infringement. Human musicians create imperfect reproductions filtered through personal interpretation and limited recall. AI systems create perfect digital copies enabling exact replication of protected elements with computational precision.14 The processing of millions of works simultaneously exceeds any individual's musical knowledge and enables reproduction of specific protected elements impossible through human memory alone. These technical processes demonstrate why AI music generation systems face substantial legal obstacles under existing copyright frameworks, creating liability for reproduction rights violations, unauthorized derivative works, and market displacement.15

2.3. Copyright Concerns Arising from AI-Generated Music

In the music industry, a significant copyright infringement liability arises for AI composition when the AI uses copyright protected work without an authority for learning and generating. There are many AI systems, such as AIVA, that create lo-fi soundtrack music, and the tracks featured in the AI Song Contest, that trained using a big data set of existing copyrighted compositions, lyrics and performances without asking permissions from the authors in the relevant jurisdictions. 16 Other AI systems train on copyrighted materials to create new "musical" outputs, which imitate styles and vocal attributes of popular artists. The generated music outputs can closely resemble copyright protected works of existing artists17.

For example, music labels sued the software startup, said to have used protected recordings of Chuck Berry and Mariah Carey without permission when it trained its AI models to replicate those original songs18.

3. Critical analysis of UMG Recordings, Inc., Capitol Recordings, LLC, Sony Music

Entertainment, Atlantic Recording Corporation, vs Suno, Inc. on the issue of fair use

3.1. Facts of the Case

This lawsuit seeks to enforce the basic principles of copyright law.19 This case concerns a

14 Id.

15 17 U.S.C. § 106(1)-(2) (2024).

16Oleksandr Bulayenko, supra note 5, at 22.

17 Id.

18 Id. at 11-12.

19 UMG Recordings, Inc. v. Suno, Inc., supra note 3, at 2 (D. Mass. June 24, 2024).

generative AI service, which allows users to generate digital music files that sound like genuine human sound recordings in response to basic inputs. 20 The Plaintiffs are record companies and recorded music businesses that together, own or exclusively control copyrights in a great majority of the most commercially valuable sound recordings in the world. Plaintiffs have a track record of embracing innovation and have entered into voluntary free-market licensing deals that authorize the use of their protected sound recordings in emerging technologies. 21 The Defendant Suno, Inc. is the company behind Suno AI, or simply Suno, a generative AI service that creates digital music files within seconds of receiving a user's prompts. Building and operating a service like Suno's requires at the outset copying and ingesting massive amounts of data to "train" a software "model" to generate outputs.

22 The plaintiffs allege several concerning practices by the defendant. Given that the foundation of its business has been to exploit copyrighted sound recordings without permission, Suno has been deliberately evasive about what exactly it has copied. 23 Suno's service trains on the expressive features of these copyrighted sound recordings for the ultimate purpose of poaching the listeners, fans, and potential licensees of the sound recordings it copied. 24 Notably Suno also claimed that its large-scale copying of sound recordings is "fair use," which was telling because fair use only arises as a defence to an otherwise unauthorized use of a copyrighted work. When Plaintiffs confronted Suno with these concessions, Suno did not respond.

25 Suno's service repeatedly generated outputs that closely matched the targeted copyrighted sound recording, which means that Suno copied those copyrighted sound recordings to include in its training data. Suno trained its model on Plaintiffs' copyrighted sound recordings. 26 The resulting harm is both immediate and far-reaching. copyrighted materials with rapid and devastating impact. Suno's service generates music with such speed and scale that it risks overrunning the market with Suno's unauthorized copying erodes the value and integrity of Plaintiffs' copyrighted sound recordings with rapid and devastating impact. The evidence demonstrates clear patterns of infringement. Suno's service generates music with such speed and scale that it risks overrunning the market with AI-generated music AI-generated content

20 Id.

21 Id. at 3.

22 Id.

23 Id.

24 Id.

25 Id. at 5.

26 Id.

and generally devaluing and substituting for human-created work.27 Furthermore, Suno also profits substantially from its infringement of Plaintiffs' copyrighted sound recordings, as the defendant's business model would not be possible without the vast troves of copyrighted sound recordings that Suno copied to train its AI model.28

3.2. Legal Issue

The main issue put forth in the lawsuit is Whether the unauthorized use of copyrighted sound recordings to train AI music generation systems constitutes copyright infringement under U.S. law, and if so, whether such use can be justified under the fair use doctrine.29This issue is further categorized in sub questions for better analysis:

1. Does copying copyrighted works into AI training datasets violate the reproduction right

under Section 106(1) of the Copyright Act?

2. Do AI-generated outputs that mimic original works constitute unauthorized derivative

works under Section 106(2)?

3. Is AI training on copyrighted works "transformative" under the first fair use factor?

4. What constitutes "market harm" when AI-generated music competes with original

recordings?

3.3. Legal Framework: Sections 106 & 107 of U.S. Copyright Law

Above given contention based on the section sec. 106 and 107 of copyright law of United States of America. Section 106 provides exclusive rights to reproduce, prepare derivative work, distribute copies, publicly perform or display their works, including digital performances of sound recordings. Whereas section 107 gives power to the people other than owner to reproduction in copies for purposes such as criticism, comment, news reporting, scholarship, or research and such practices do not constitute infringement of copyright.

27 UMG Recordings, Inc. v. Suno, Inc., supra note 3, at 5 (D. Mass. June 24, 2024).

28 Id.

29 Id. at 25

4. Author’s Analysis and Opinion

The main issue put forth in the lawsuit is Whether the unauthorized use of copyrighted sound recordings to train AI music generation systems constitutes copyright infringement under U. S. law, and if so, whether such use can be justified under the fair use doctrine. 30This issue is further categorized in sub questions for better analysis: 4. 1. Does copying copyrighted works into AI training datasets violate the reproduction right under Section 106(1) of the Copyright Act? The copying of copyrighted works into AI training datasets implicates the reproduction right under Section 106(1) of the Copyright Act, as expounded upon in the Copyright and artificial intelligence Part 3: Generative AI Training pre-publication version.

The reproduction right extends to the protected elements of the copyrighted work, in whole or in part, and to non-literal copies with significant variations, as long they are substantially similar to the original. 31 As discussed above, the use of those works in preparing a training dataset and training a model implicates the reproduction right, but copying the resulting weights will constitute infringement only where there is substantial similarity. 32 Whether a model's weights implicate the reproduction or derivative work rights turns on whether the model has retained or memorized substantial protectable expression from the work(s) at issue. 33 This distinction is particularly relevant because generative AI models sometimes output material that replicates or closely resembles copyrighted works.

Users have demonstrated that generative AI can produce near exact replicas of still images from movies, copyrightable characters, or text from news stories. Such outputs likely infringe the reproduction right and, to the extent they adapt the originals, the right to prepare derivative works. 34 In MAI Systems Corp. v. Peak Computer, Inc. , 991 F. 2d 511 (9th Cir. 1993), held that loading copyrighted software into RAM constitutes a reproduction under Section 106(1), as the copy is sufficiently fixed to be perceived and used, because after loading the software, it was possible to view information regarding the computer’s status, the copy in RAM met the ‘sufficiently

30 UMG Recordings, Inc. v. Suno, Inc., supra note 31, at 25 (Joint Statement).

31 17 U.S.C. § 106(1). See Feist Publ'ns, Inc. v. Rural Tel. Serv. Co., 499 U.S. 340, 361 (1991).

32 U.S. COPYRIGHT OFFICE, COPYRIGHT AND ARTIFICIAL INTELLIGENCE, PART 3: GENERATIVE

AI TRAINING, at 30 (pre-publication version, May 2025), https://copyright.gov/policy/artificial-intelligence/.

33 UMG Recordings, Inc. v. Suno, Inc., supra note 31, at 30 (Joint Statement).

34 Id. at 31.

permanent or stable’ requirement of the Copyright Act. 35 This interpretation supports the view that copying copyrighted works into memory during AI training similarly implicates the reproduction right. Similarly, in Perfect 10, Inc. v. Amazon. com, Inc. , the Ninth Circuit reaffirmed that copying copyrighted images onto servers to generate thumbnails for a search engine constituted direct infringement of the reproduction and display rights. The court emphasized that such automated or computational copying does not exempt the use from liability, even if the purpose is functional or transformative: “A photographic image is a work that is ‘fixed’ ... when embodied (i. e. , stored) in a computer’s server ... The image stored in the computer is the ‘copy’ of the work for purposes of copyright law.

”36 This principle applies directly to AI training, where copyrighted works are copied and stored on servers for computational processing. Thus, AI developers who copy works into memory for training purposes risk infringing reproduction rights—unless a valid defence such as fair use applies. Based on established precedent from MAI Systems and Perfect 10, the unauthorized copying of copyrighted works during AI training satisfies the reproduction element of a copyright infringement claim. The copying required for dataset creation clearly engages the reproduction right under existing jurisprudence, as it meets both the fixation and substantial similarity requirements necessary to establish prima facie infringement.

4.2. Do AI-generated outputs that mimic original works constitute unauthorized

derivative works under Section 106(2)? Under 17 U.S.C. § 106(2), a derivative work is one based on one or more preexisting works and recasts, transforms, or adapts the original in a way that incorporates its protected expressive elements. Courts have interpreted this to mean that substantial similarity is a key factor in determining whether a new work unlawfully derives from an original. A critical distinction emerges from recent judicial interpretations, particularly the contrasting

outcomes in Kadrey v. Meta37 and Andersen v. Stability AI. While the training process itself implicates the reproduction right through the downloading, transferring, and converting of

35 MAI Sys. Corp. v. Peak Comput., Inc., 991 F.2d 511, 518 (9th Cir. 1993).

36 Perfect 10, Inc. v. Amazon.com, Inc., 508 F.3d 1146, 1160–61 (9th Cir. 2007).

37 Kadrey v. Meta Platforms, Inc., No. 23-cv-3417, 2023 WL 8039640, at 1 (N.D. Cal. Nov. 20, 2023).

copyrighted works across storage mediums, the resulting AI model only constitutes derivative work infringement where there is substantial similarity to the original work. 38 The Andersen court's refusal to dismiss claims was predicated on "sufficient allegations that copies or protected elements remained, in some format, within the model,"39 distinguishing cases where models have "retained or memorized substantial protectable expression" from those involving mere pattern recognition. In Lewis Galoob Toys, Inc. v. Nintendo of America, Inc. , the Ninth Circuit clearly articulated this threshold: “A derivative work must incorporate a protected work in some concrete or permanent form.

”40 The court distinguished between temporary, user-controlled alterations (like those created by the Game Genie device) and actual derivative works that embody original protected expression in a fixed, tangible medium. Because the Game Genie merely modified gameplay visuals in real time without storing or embedding copyrighted content, it did not create a derivative work. This rule helps courts decide AI music cases. If an AI system creates music that borrows protected parts of existing songs - like memorable melodies or distinctive rhythms - and that music exists in a permanent form (like a digital file or recording), it could be an illegal derivative work. When discussing whether AI-generated music that resembles original songs infringes copyright, the reasoning in Bright Tunes Music Corp. v. Harrisongs Music, Ltd. is particularly instructive.

There, the court held that subconscious or unintentional copying can constitute infringement where the new work is substantially similar to the original. The decision emphasized that: “The Court ... addressed the question as to whether the copying of music had to be intentional. It stated that it did not. ... Circumstantial evidence was just as good. ”41 This principle suggests that AI-generated content, even absent direct intent or human authorship, may infringe copyright if the output reflects substantial similarity to protected works. Consequently, AI systems that reproduce protected expressive elements and fix them in tangible media create unauthorized derivative works, with the legal framework focusing on substantiality of copied expression rather than technological method.

38 U.S. COPYRIGHT OFFICE, supra note 35, at 27.

39 Andersen v. Stability AI Ltd., 744 F. Supp. 3d 956, 982–84 (N.D. Cal. 2024).

40 Lewis Galoob Toys, Inc. v. Nintendo of Am., Inc., 964 F.2d 965, 969 (9th Cir. 1992).

41 Bright Tunes Music Corp. v. Harrisongs Music, Ltd., 420 F. Supp. 177 (S.D.N.Y. 1976).

4.3. Is AI training on copyrighted works "transformative" under the first Fair Use

factor? Under fair use law, "transformative" use means adding new meaning to original work or using it for a completely different purpose. The Supreme Court's Warhol decision made this stricter - it's not enough to just change something. You have to use it for a fundamentally different reason than the original. In the Warhol case, the Supreme Court said that simply changing or adding meaning isn't sufficient. The Court noted that courts must consider commercial factors too. If a new work serves basically the same purpose as the original and makes money from it, fair use probably doesn't apply without a really good justification. 42 This creates problems for AI companies like Suno. They use copyrighted music to train systems that create new music for commercial services.

Even though the AI output might sound different, the company is still using copyrighted works commercially to make competing music products. This isn't clearly transformative like parody, criticism, or education would be. The Supreme Court warned that letting commercial uses like this count as "transformative" would essentially eliminate copyright owners' rights to control derivative works. 43 So unless AI training serves a genuinely different public purpose, it probably won't qualify as transformative use. The U. S. Copyright Office has articulated a nuanced framework for evaluating the transformative nature of AI training under the first fair use factor, emphasizing that analysis must extend beyond the immediate act of copying to encompass the ultimate purpose and deployment of AI systems.

The Office explicitly states that "different uses during AI development and deployment require separate consideration,"44 rejecting a monolithic approach to fair use analysis in the AI context. While acknowledging that "training a generative AI foundation model on a large and diverse dataset will often be transformative,"45 the Office draws critical distinctions based on the model's intended application. Specifically, the Office views uses as less transformative when models are "trained on specific types of works in order to produce content that shares the purpose of appealing to a particular audience," citing the example that "training an audio model on sound recordings for deployment in a system to generate new sound recordings aims to occupy the same space in the market for music and

42 Andy Warhol Found. for the Visual Arts, Inc. v. Goldsmith, 143 S. Ct. 1258, 1275–76 (2023).

43 Id. at 1274.

44 U.S. COPYRIGHT OFFICE, supra note 35, at 36.

45 Id. at 45.

satisfy the same consumer desire for entertainment and enjoyment. "46 Conversely, the same model deployed for different purposes, such as "removing unwanted distortion from sound recordings," would be considered more transformative. 47 The Office explicitly rejects two prevalent arguments advanced by AI developers: first, that AI training constitutes purely "nonexpressive48" use when "the resulting model is used to generate expressive content, or potentially reproduce copyrighted expression," and second, that AI learning parallels human learning, the office noted that "AI learning is different from human learning in ways that are material to the copyright analysis" because "generative AI training involves the creation of perfect copies with the ability to analyse works nearly instantaneously.

"49 This framework reflects the Office's recognition that generative AI models "may simultaneously serve transformative and non-transformative purposes,"7 necessitating a case-by-case analysis that considers both the training methodology and the ultimate deployment context. The U. S. copyright Office's framework aligns closely with Warhol's restrictive approach to transformative use, requiring AI training to serve fundamentally different purposes from the original copyrighted works. When AI companies train models on copyrighted content to generate competing commercial products, like Suno's music generation, they face the same analytical hurdle that defeated the Warhol Foundation: commercial use for substantially similar purposes without compelling justification.

Under this combined framework, most commercial AI training operations will likely fail the transformative use analysis unless they can demonstrate deployment for genuinely non-substitutive purposes. 4. 4. What constitutes "market harm" when AI-generated music competes with original recordings? Under the fourth fair use factor, the U. S. Copyright Office identifies “market harm” as encompassing several distinct but overlapping concerns, especially in the context of AIgenerated music. Most directly, market harm arises when “outputs substitute for the original” such that they “result in lost sales or displace demand” for the copyrighted works used in training. 50 The risk is heightened when generative AI systems produce “near exact replicas” of sound recordings or closely mimic the style and voice of original artists—outputs that “likely

46 U.S. COPYRIGHT OFFICE, supra note 35, at 46.

47 Id.

48 Id. at 48.

49 Id.

50 Id. at 73.

infringe the reproduction right and, to the extent they adapt the originals, the right to prepare derivative works. ”51 Beyond substitution, the Office acknowledges the threat of “market dilution,” where even non-identical outputs may “undermine the value of creative works” by flooding the market with synthetic alternatives. 52 Critically, the Office emphasizes that harm also includes the loss of “licensing opportunities,” noting that “unauthorized use of copyrighted works in training can cause significant harm by undermining or bypassing emerging licensing markets. ”53 In the realm of sound recordings—where licensing structures are already wellestablished—such unlicensed use is viewed as particularly injurious.

As the Office cautions, “[t]he development of a licensing market for generative AI is relevant to the fair use analysis” because its circumvention by AI developers “could deprive copyright owners of actual or potential revenues. ”54 This comprehensive framing of market harm strongly weighs against a finding of fair use when generative AI music competes directly or indirectly with original copyrighted recordings. Under the fourth fair use factor, courts assess whether the challenged use harms the actual or potential market for the original or its derivatives. In Harper & Row Publishers, Inc. v.

Nation Enterprises, the Supreme Court emphasized that market harm includes not only actual financial loss but also interference with potential licensing markets, holding that even prospective displacement is relevant if the use “would adversely affect the potential market”55 This expansive view applies directly to AI-generated music, which may substitute for both present streaming revenue and future AI licensing streams. In Campbell v. Acuff-Rose Music, Inc. , the Court further refined this by stating that the central question is whether the new work serves as a “market substitute” for the original or harms the market for authorized derivative uses. 56 AI-generated outputs that closely mimic copyrighted music—especially when consumed as entertainment—pose a cognizable substitution risk, reducing both demand for originals and licensing incentives.

The convergence of regulatory guidance and Supreme Court precedent establishes that AIgenerated music faces formidable obstacles under the fourth fair use factor through multiple

51 U.S. COPYRIGHT OFFICE, supra note 35, at 31.

52 Id. at 64.

53 Id. at 61.

54 Id. at 63.

55 Harper & Row, Publ’rs, Inc. v. Nation Enters., 471 U.S. 539, 568 (1985).

56 Campbell v. Acuff-Rose Music, Inc., 510 U.S. 569, 590–91 (1994).

layers of cognizable market harm. Cases like UMG v. Suno exemplify how AI platforms create immediate substitution effects by enabling rapid generation of music that directly competes with copyrighted recordings in the entertainment market, while simultaneously circumventing established licensing ecosystems that represent primary revenue sources for rights holders. The technological capability of AI systems to flood markets with competing content virtually instantaneously transforms market impact from incremental harm into systemic disruption, creating an existential threat to copyright-dependent business models.

Consequently, when AI systems produce music serving identical entertainment purposes as copyrighted works— regardless of technical differences—they risk displacing both current revenue streams and future licensing opportunities that courts have consistently protected, creating a strong presumption against fair use defences for commercial AI music generation.

4.5. AI Authorship in Musical Works: Comparative Analysis of Indian and U.S. Legal

Frameworks Having examined various copyright violations occurring when AI creates music, this section analyses the critical question of authorship for musical works created by humans with AI assistance, comparing Indian and U. S. legal approaches. Indian copyright law operates under the "sweat of the brow" principle, prioritizing effort, labour, and investment over pure originality, establishing a relatively low threshold for copyright protection. 57 This contrasts with U. S. jurisprudence, which applies a higher "modicum of creativity" standard from Feist Publications, Inc. v. Rural Telephone Service Co. 58 Despite these philosophical differences, both systems require human authorship for copyright protection, excluding entirely computer-generated works unless legislative intervention occurs.

The Indian Copyright Act explicitly addresses this through Section 2(d)(vi), defining "author" for computer-generated works as "the person who causes the work to be created. "59 The Supreme Court in Indian Performing Rights Society v. Eastern India Motion Picture Association clarified that for musical works, "the composer alone has copyright," though suggesting broader protection might be warranted. 60 The recent Ankit Sahani case illustrates these complexities, where digitally created art using AI software 57 Shreya Matilal, The Indian "Sweat of the Brow" Doctrine in Copyright Law, 3 NUJS L. REV. 259, 261-63 (2010).

58 Feist Publ'ns, Inc. v. Rural Tel. Serv. Co., 499 U.S. 340, 345-46 (1991).

59 The Copyright Act, 1957, § 2(d)(vi) (India).

60 Indian Performing Rights Soc'y v. E. India Motion Picture Ass'n, A.I.R. 1977 S.C. 1443, 1448 (India).

"RAGHAV" faced rejection by the U. S. Copyright Office for lack of human creativity, despite the artist's creative prompts and instructions. 61 While this artwork received joint authorship status in Canada, Indian courts have yet to definitively rule on AI-assisted creative works' originality and human effort requirements. This divergence highlights the urgent need for judicial clarity on AI involvement in musical and artistic creation, as current frameworks struggle with hybrid human-AI authorship scenarios that increasingly define contemporary creative practices. Notably, while the United States actively develops comprehensive AI copyright frameworks through detailed regulatory guidance and landmark litigation like UMG v. Suno, Indian copyright law remains underdeveloped in addressing AI-specific infringement scenarios.

India urgently requires legislative reform or judicial interpretation to match the evolving demands of AI-driven creative industries.

5. Conclusion

The UMG v. Suno litigation based upon the fundamental tensions between artificial intelligence capabilities and copyright protection, revealing critical inadequacies in existing legal frameworks when applied to AI music generation. This analysis of the four central legal questions demonstrates that Suno faces substantial liability under current copyright doctrine. The copying of copyrighted works during AI training clearly violates Section 106(1) reproduction rights. Established precedent from MAI Systems and Perfect 10 confirms that loading copyrighted material into computer memory constitutes actionable reproduction, directly applicable to AI training processes that require persistent storage and computational processing of millions of copyrighted recordings.

Further the AI-generated outputs that incorporate substantial similarity to original works constitute unauthorized derivative works under Section 106(2). The Bright Tunes decision establishes that even unintentional copying can constitute infringement where substantial similarity exists, making Suno's outputs legally vulnerable regardless of algorithmic intent when they reproduce protected expressive elements. The Supreme Court's restrictive Warhol framework, Suno's commercial music generation fails the transformative use test. The platform trains on copyrighted music to generate competing entertainment content for substantially similar commercial purposes—precisely the type of use 61 Vedika Chawla, Ankit Sahni's AI "Co-authored" Artwork Denied Registration by US, Continues to be Registered in India, SPICY IP (Dec. 15, 2023), https://spicyip.

com/2023/12/ankit-sahnis-ai-co-authored-artwork-deniedregistration-by-us-continues-to-be-registered-in-india. html.

Warhol warns could "swallow the copyright owner's exclusive right to prepare derivative works. " The U. S. Copyright Office's guidance reinforces this conclusion, explicitly stating that training audio models to generate new sound recordings "aims to occupy the same space in the market for music. "The fourth fair use factor weighs decisively against Suno through multiple forms of cognizable harm. AI-generated music creates immediate substitution effects by competing directly with original recordings in entertainment markets while circumventing established licensing structures that represent primary revenue sources for rights holders. This displacement threatens both current streaming revenue and future licensing opportunities that courts have consistently protected under Harper & Row and Campbell.

When we put all these legal tests together, AI music companies face major hurdles under current law. AI has creative potential, but the legal deck is stacked in favour of established music rights holders over tech developers. Without licensing agreements or new laws, AI companies will likely lose these cases. The core issue is that fair use doctrine wasn't designed for technology that can analyse vast amounts of music at lightning speed. We need legal reforms that protect both technological progress and creators' livelihoods.