IMPACT OF AI ON COPYRIGHT LAW: WHO OWNS AI-GENERATED WORKS? BY - SIMRAN GAHLOT

IMPACT OF AI ON COPYRIGHT LAW: WHO OWNS AI-GENERATED WORKS?
 
AUTHORED BY - SIMRAN GAHLOT
 
 

Abstract

The rapid advancements in Artificial Intelligence (AI) have significantly disrupted copyright law, particularly in determining authorship and ownership of AI-generated works. Traditional copyright frameworks, which emphasize human creativity and intellectual effort, struggle to accommodate AI’s role as a creator. This paper examines the challenges posed by AI-generated works within the Indian Copyright Act, 1957, analyzing whether existing legal provisions sufficiently address issues of originality, attribution, and liability. It explores global legal perspectives, including the U.S. fair use doctrine and the EU’s Text and Data Mining (TDM) exceptions, to draw insights for India. The study further discusses ethical dilemmas, such as copyright infringement concerns arising from AI training on protected content and the potential for AI to be recognized as a legal entity. By assessing landmark case laws, including Naruto v. Slater,[1] Thaler v. USPTO[2], and Getty Images v. Stability AI[3], this paper highlights the evolving discourse surrounding AI-generated works. It ultimately calls for legislative clarity and possible regulatory reforms to balance innovation with copyright protection in an AI-driven creative landscape.
 
Keywords: AI-generated works, copyright law, originality, authorship, fair use, AI liability, intellectual property rights, AI-generated music, legal framework, ethical concerns, copyright infringement, AI training data, Indian Copyright Act, AI regulation, global legal approaches.
 

Introduction

The rapid advancements in Artificial Intelligence (AI) over the past two years have significantly reshaped industries, particularly in the creative and intellectual property domains. AI has evolved from being a mere tool in the hands of its creator to becoming a creator itself. This unprecedented growth has sparked a global debate on the adequacy of existing legal frameworks, as nations struggle to keep pace with the evolving technological landscape.
 
In the realm of copyright law, AI challenges the traditional notions of originality and authorship. Historically, copyright protection has been granted based on human creativity and intellectual effort. However, with AI now capable of composing music, generating literature, and producing art, the question arises—who owns AI-generated works? Can merely submitting a prompt or instruction to an AI model qualify as authorship under copyright law?
The Indian Copyright Act, 1957, does not explicitly recognize AI as an author, creating uncertainty regarding ownership and protection of AI-generated works. As AI tools like ChatGPT continue to redefine content creation, the absence of clear legal provisions raises complex issues about attribution and rights. This paper seeks to analyze whether the existing copyright framework in India is equipped to handle these emerging challenges, with a specific focus on AI’s role in music composition and production. By examining legal perspectives and recent technological advancements, this study aims to explore the evolving discourse surrounding AI-generated works and their implications on copyright law in India.
 
Navigating the Copyright Act: Implications foe AI-Generated Content
Authorship in the Age of AI
definition of “author” as any person who causes the work to be generated by a computer, thus eliminating any chances of machines getting authorship of the work developed by it, independent of any human interference[4].
 
In the evolving landscape of AI-generated content, the question of authorship and ownership presents a significant legal challenge. Under the current framework, none of the involved parties—the creator of the AI system, the AI itself, or even the copyright holders of the data used to train the AI—seem to have a clear claim to legal protection. Instead, the individual who simply provides a prompt or instruction for the AI to generate music or any other creative work may be the only one entitled to protection. However, this raises concerns about whether such minimal human input should be sufficient to establish authorship.
 
An alternative approach could be to compare the relationship between AI and its user to that of an employer and an employee, where the output is treated as a work-for-hire. However, this analogy loses validity, as AI lacks the legal capacity to enter into contracts or consent to such relationships. Furthermore, AI developers may argue for co-authorship, asserting that their contribution in designing and programming the system, particularly during the conceptualization phase, plays a fundamental role in the creative process. This could lead to disputes over ownership rights, especially in cases where AI-generated works attain significant commercial value.
 
Another critical factor in determining authorship is distinguishing between AI-assisted and AI-generated works. In AI-assisted works, human creators utilize AI tools as aids, but their creative input remains central to the final output. In contrast, AI-generated works involve minimal or no human intervention beyond providing an initial instruction. Indian copyright law currently recognizes and protects only AI-assisted works, as they still embody human creativity. However, fully AI-generated works fall into a legal gray area, raising questions about whether copyright law needs to evolve to address these emerging challenges.
 
This complexity underscores the urgent need for legislative clarity to address the implications of AI-generated content, ensuring that the legal framework remains equipped to balance innovation, creative rights, and fair attribution in the age of artificial intelligence.
 
The Need for Originality and Creativity
For a musical work to be eligible for copyright protection in India, it must be expressed in a tangible form, as copyright law does not safeguard mere ideas.
 
"originality" as a fundamental requirement for a work to qualify for protection.[5] However, the term "original work" is not explicitly defined within the statute. Courts typically assess originality by examining the relationship between an idea and its expression, often invoking the Doctrine of Merger[6]. This evaluation focuses on whether the work reflects the author’s skill, effort, and creative input. Judicial interpretations differentiate between works that merely involve labour and those that require both skill and judgment.
 
AI-generated works pose a unique challenge in this context, as AI operates through algorithms and data processing rather than human creativity, intuition, or judgment. While AI can generate compositions resembling human-created music, its process does not align with conventional concepts of authorship or creative labour. Indian courts have interpreted originality in a manner similar to U.S. courts, which apply the "Modicum of Creativity[7]" standard. This principle holds that only those works demonstrating a sufficient degree of skill and judgment meet the originality threshold. Given AI's evolving role as a creator rather than just an assistive tool, courts may face significant challenges in reconciling traditional copyright principles with the distinct nature of AI-generated works.
 

Ethical Dilemmas and Copyright Infringement

AI operates on the basis of large language models, processing vast amounts of data using advanced algorithms. In traditional cases involving the remix or adaptation of a copyrighted song or musical work, an individual must obtain permission from the rightful owner before making modifications or using the work. This requirement aligns with Section 52(1)(j)[8], which governs such uses.
 
However, in the case of AI-generated music, no such permissions are sought. The developers of AI models do not explicitly obtain consent before training AI on copyrighted content, nor does the AI itself have the capability to do so. This raises significant concerns about liability—who is responsible when AI incorporates copyrighted material from its training data? If an AI-generated composition inadvertently includes elements of a protected work, it remains uncertain whether the responsibility lies with the AI's developer, the user, or another entity. The legal framework currently lacks clarity on how liability should be assigned in such cases, creating complex ethical and legal challenges in the realm of copyright law.
 

Recognizing AI as a Separate Entity

The idea of granting AI a distinct legal identity has been proposed as a possible solution to the challenges surrounding AI-generated works. However, this approach raises complex legal, ethical, and practical concerns, particularly regarding authorship, ownership, accountability, and moral rights.
 
One of the fundamental obstacles is that AI lacks legal agency, meaning it cannot enter into contracts, exercise rights, or be held accountable for its actions in the same way as human creators. Under Section 57[9], authors are granted moral rights, including the right to paternity (the right to be recognized as the author) and the right to integrity (the right to object to distortion or modification of their work). In the case of AI-generated content, enforcing these rights becomes problematic, as AI lacks personal identity, reputation, or the ability to assert claims over its creations.
 
Furthermore, the Act requires authors to receive royalties for the use of their works, ensuring they are compensated for their intellectual efforts. However, AI, being a machine, does not have the capacity to claim, manage, or determine royalties. This raises questions about whether AI-generated works should be assigned to its developer, user, or another entity, and how financial benefits from such works should be distributed.
 
Another major concern is liability for objectionable or infringing content. If an AI creates content that violates copyright laws, spreads misinformation, or produces harmful material, determining who should be held accountable—the AI’s developer, the user, or another party—becomes a significant legal challenge. Unlike human creators, AI lacks moral judgment, ethical reasoning, or intent, making traditional liability frameworks difficult to apply.
 
While recognizing AI as an author could introduce new dimensions to copyright law, it also presents profound challenges that must be carefully considered. Establishing a legal framework for AI-generated works would require clear guidelines on ownership, liability, and ethical responsibility to ensure fairness and accountability while fostering innovation in AI-driven creativity.
 

Charting a Path Forward for Copyright Law

As artificial intelligence (AI) continues to redefine creative production, lawmakers must determine how copyright law should adapt to address AI-generated works. One possible approach is to exclude AI-generated works from copyright protection if they are produced without human involvement and are openly accessible to all. The decision to grant protection should depend on the costs and benefits associated with copyright enforcement. If a work is entirely generated by AI, it should not grant any individual the right to restrict its reproduction or distribution. However, in cases where a human has played a significant role in the creation process, courts should assess copyright claims on an ad-hoc basis to evaluate the extent of human contribution.
 

Learning from Global Approaches

Given that the challenges posed by AI-generated content are global, Indian lawmakers could look beyond national boundaries and analyze how other jurisdictions, such as the United States (US) and the European Union (EU), are tackling similar issues.
1.      Text and Data Mining (TDM) Exceptions
a.       The EU’s copyright framework recognizes TDM exceptions, which allow automated analysis of large datasets to identify trends and generate insights.
b.      This exception applies only in legally defined scenarios, ensuring that authors retain some control over their works.
c.       In some cases, copyright holders can opt-out of TDM-based usage, making it a passive permission system rather than a blanket authorization.
2.      Fair Use Doctrine (US)
a.       The US fair use doctrine allows individuals to use copyrighted material without obtaining prior consent from the copyright owner under specific conditions.
b.      Whether a work qualifies under fair use depends on several factors, including:
                                                              i.      The purpose and character of the use (e.g., commercial or educational).
                                                            ii.      The nature of the copyrighted work.
                                                          iii.      The amount and substantiality of the portion used.
                                                          iv.      The effect of the use on the market for the original work.
c.       Notably, both the TDM exception and fair use have been applied to cases involving scientific research, where authors' permission is not required for data utilization.
3.      AI-Specific Legal Frameworks
a.       The EU is in the final stages of passing the AI Act, which aims to regulate AI technology and its impact across various sectors, including intellectual property rights.
b.      India could consider implementing a sui generis system to provide tailored intellectual property (IP) protections for AI-generated content.
c.       A specialized legal framework addressing ownership, liability, and ethical concerns arising from AI-generated works would help bridge the gaps in the existing copyright laws.
 

Technological Solutions to AI-Generated Copyright Issues

Aside from legislative reforms, AI developers could implement technological measures to ensure better compliance with copyright norms. Some potential solutions include:
1.      Audio Steganography and Digital Watermarking
a.       Audio steganography is a technique for embedding hidden information within audio files.
b.      Developers could integrate digital watermarks into AI-generated music or creative works, ensuring that the source of the content is traceable.
2.      AI-Generated Citations
a.       AI models could be designed to cite their data sources whenever generating content derived from existing works.
b.      This could function similarly to academic referencing, helping users distinguish original AI-generated content from content based on prior copyrighted works.
3.      Fair Use Analysis for AI Training Data
a.       Lawmakers should clarify whether storing and using copyrighted works in AI training databases constitutes fair use under copyright law.
b.      AI developers should implement transparent policies regarding the use of copyrighted content in training datasets.
 
Relevant Case Laws
AI-Generated Works and Copyright Protection
Case 1: Naruto v. Slater – Monkey Selfie Case[10]
Key Issue: Whether a non-human entity (a monkey) can be recognized as the author of a copyrighted work.
Facts:
·         A macaque monkey, Naruto, took a selfie using a camera owned by photographer David Slater.
·         Slater claimed copyright, but PETA argued that Naruto should own the copyright.
·         The court ruled that non-human entities cannot hold copyright under U.S. law.
Relevance to AI:
·         If a monkey cannot be the author of a copyrighted work, can AI?
·         This case sets a precedent against AI being recognized as an author under copyright law.

Case 2: Thaler v. USPTO – AI as an Inventor[11]

Key Issue: Can AI be named as an inventor or creator under patent and copyright laws?
Facts:
·         Dr. Stephen Thaler filed a patent application naming his AI system, DABUS, as the inventor.
·         The USPTO rejected the application, ruling that only humans can be inventors under U.S. patent law.
·         The Federal Circuit upheld the ruling, stating that AI lacks the legal status of a "person" required for authorship.
Relevance to AI Copyright:
·         If AI cannot be an inventor, it likely cannot be an author under copyright law.
·         Courts emphasize human creativity and intellectual effort in granting copyright.

Originality and the “Modicum of Creativity” Standard

Case 3: Feist Publications, Inc. v. Rural Telephone Service Co.[12]

Key Issue: What qualifies as "original work" under copyright law?
Facts:
·         Feist Publications used telephone directory listings from Rural Telephone Service.
·         The Supreme Court ruled that the mere collection of facts is not copyrightable.
·         For a work to be copyrightable, it must have a "modicum of creativity."
Relevance to AI:
·         AI arranges and processes data but does not exhibit human creativity.
·         Could AI-generated works fail to meet the originality standard?

Indian Case Laws on Copyright and AI Implications

Case 4: Eastern Book Company v. D.B. Modak [13]

Key Issue: Whether manual selection and arrangement of text qualifies as an original work under Indian copyright law.
Facts:
·         Eastern Book Company (EBC) compiled Supreme Court judgments with editorial notes.
·         The court ruled that mere compilation is not enough; it must involve skill and creativity.
Relevance to AI:
·         AI-generated works may lack skill and judgment, making them ineligible for copyright.

Case 5: R.G. Anand v. Delux Films [14]

Key Issue: When does copying a concept constitute copyright infringement?
Facts:
·         The Supreme Court ruled that ideas are not protected under copyright—only their specific expression is.
Relevance to AI:
·         AI reproduces ideas in different forms, which may not qualify as copyright infringement under Indian law.

AI and Copyright Infringement

Case 6: Authors Guild v. Google, Inc.[15] - Fair Use in AI Training

Key Issue: Does scanning and digitizing books for AI-based searches infringe copyright?
Facts:
·         Google digitized millions of books for its Google Books project, allowing AI-driven search functionalities.
·         The court ruled this was fair use because it was transformative and did not replace the original works.
Relevance to AI:
·         AI models like ChatGPT, MidJourney, and DALL·E train on copyrighted works.
·         Should such training be considered fair use or copyright infringement?

Case 7: Getty Images v. Stability AI Ltd.[16]

Key Issue: Stability AI (maker of Stable Diffusion) was sued for using copyrighted images without permission in training its AI model.
Facts:
·         Getty Images accused Stability AI of scraping millions of images without licensing them.
·         The lawsuit argued that AI-generated images were derivative works of copyrighted materials.
Relevance to AI:
·         Courts are currently deciding whether training AI on copyrighted content without permission is legal.
·         The ruling will impact AI-generated music, literature, and visual art.
 
Conclusion

As artificial intelligence continues to advance, it challenges traditional notions of authorship, originality, and liability in copyright law. Existing legal frameworks, including the Indian Copyright Act, 1957, remain inadequate in addressing the complexities of AI-generated works. While human creativity has historically been central to copyright protection, AI-generated content raises fundamental questions about ownership—should protection be granted to the developer, the user, or the AI itself?

Global legal precedents, including Naruto v. Slater, Thaler v. USPTO, and Getty Images v. Stability AI, highlight courts' reluctance to recognize non-human entities as authors. The "modicum of creativity" standard, as reaffirmed in Feist Publications, Inc. v. Rural Telephone Service Co., further complicates the copyrightability of AI-generated works, which may lack the human intellectual effort courts traditionally require.
 
Ethical dilemmas surrounding AI training on copyrighted material without consent underscore the need for clearer legal frameworks. While the U.S. Fair Use doctrine and the EU’s Text and Data Mining (TDM) exceptions offer some guidance, India must develop its own approach—whether through adapting existing laws or creating a sui generis system tailored to AI.
 
To strike a balance between innovation and copyright protection, a multi-faceted approach is necessary. Legislative reforms, judicial clarity, and technological solutions such as digital watermarking and AI-generated citations can help address concerns related to originality, infringement, and attribution. By proactively addressing these challenges, India can foster an AI-driven creative landscape while ensuring fair recognition and protection for human creators.

[1] 888 F.3d 418 (9th Cir. 2018)
[2] 43 F.4th 1207 (Fed. Cir. 2022)
[3] No. 1:23-cv-00135 (D. Del. filed Feb. 3, 2023)
[4] The Copyright Act, No. 14 of 1957, § 2(d), India Code (1957)
[5] The Copyright Act, No. 14 of 1957, § 13, India Code (1957)
[6] Baker v. Selden, 101 U.S. 99 (1879)
[7] Feist Publications, Inc. v. Rural Telephone Service Co., 499 U.S. 340 (1991)
[8] Indian Copyright Act, 1957
[9] The Copyright Act,1957
[10] 888 F.3d 418 (9th Cir. 2018)
[11] 43 F.4th 1207, 1209 (Fed. Cir. 2022)
[12] 499 U.S. 340 (1991)
[13] (2008) 1 SCC 1 (India)
[14] (1978) 4 SCC 118 (India)
[15] 804 F.3d 202 (2d Cir. 2015)
[16] [2023] EWHC 752 (Ch)