AI And Authorship Debate From A Copyright Perspective
Author: LawSphereX
Category: Technology, Cyber & Privacy
Published:
Updated:
AI AND AUTHORSHIP DEBATE FROM A COPYRIGHT
PERSPECTIVE
Vineet Bidhuri, LL.M., Hidayatullah National Law University
ABSTRACT
The rapid growth of artificial intelligence has deeply impacted the lives of artists and creators. For instance, the creation of texts, images, music, and other works of art by artificial intelligence systems raises relevant and important legal questions regarding authorship and ownership in the context of existing copyright laws. This paper aims to study the changing dynamics between artificial intelligence and copyright laws, with specific reference to the Indian legal system. It discusses the concept of ‘author’ in the Copyright Act, 1957, and assess the feasibility of incorporating artificial intelligence works within the existing copyright laws. Additionally, it sheds light on the complexity associated with the question of authorship in AI-generated works.
It also highlights various stakeholders involved in the creation of AIgenerated works, including authors, developers, users, and data providers, in relation to the claims of authorship over AI-generated works. It further discusses various contemporary legal issues associated with AI-generated works, including training data, copyright infringement, and personality rights. Finally, the paper tries to recommend an effective balance between human creativity and artificial intelligence in contemporary times. It is apparent that the existing Indian copyright laws do not address the issues relating to AI-generated works. Therefore, it is important that an effective balance between human creativity and artificial intelligence is achieved.
INTRODUCTION
The rapid development of AI technology has significantly changed our understanding and the way we look at creative work. In our daily lives, we consume so much content that is created or in some way influenced by AI, sometimes making it difficult to differentiate it from the original. Currently, AI technology is capable of creating literary works, visual arts, music, and other creative works with minimal or no human intervention. However, this emerging reality has raised significant and fundamental issues regarding the current relevance of the Copyright Act in relation to AI technology and creative work. The Copyright Act, 1957 was amended in 1994, adding Section 2(o) defining “literary work” to include ‘computer programmes, tables and compilations including computer databases’ to address the developing nature of technology.
According to Section 2(d)(vi) of the Copyright Act, “for a computer-generated literary, dramatic, musical, or artistic work, the author is the person by whom the work is caused to be created. ” The assumption was that even with the help of computers and other technology, the creative role and intervention of humans would always remain relevant and notable. 1 However, the growth and rapid development of AI technology have challenged this assumption and raised fundamental issues relating to the applicability and relevance of the Copyright Act in relation to AI technology and creative work. Currently, AI technology is capable of creating literary works, visual arts, music, and other works with minimal or no human intervention. Therefore, the question arises: how do we identify the “author” under the Copyright Act?
Copyright is an intellectual property that protects the expression of original ideas and creativity. According to Section 14 of the Copyright Act, “the author of a literary, dramatic, musical or artistic work has the exclusive right to reproduce the work in any material form, to publish the work, to adapt the work, and to communicate the work to the public. ” According to Section 17 of the Copyright Act, “the author of a literary or dramatic work is the first owner of the copyright in the work. ”2 Considering the fact that the role of AI is becoming crucial for the development of the economy as well as the field of technology, the Indian government has introduced various policy initiatives such as the “AI for All” plan. This shows that the intellectual property law framework could be subject to revision as the role of AI increases in the development of intellectual works.
1 Copyright Act, 1957, § 2(d)(vi) (India).
2 Copyright Act, 1957, §§ 14, 17 (India).
Keeping the above background in mind, the research paper tries to address the debate on the ownership rights for the works created by AI, along with the authorship rights that can be attributed to these works in the framework of Indian copyright law.
LITERATURE REVIEW
Journal Articles and Books
1. Naithani, P. (2022). “Issues of Authorship and Ownership in Work Created by Artificial
Intelligence — Indian Copyright Law Perspective.” NTUT Journal of Intellectual Property Law & Management, 11(1), 7. Naithani (2022) examines the problems of authorship and ownership under Indian copyright law in the context of AI-generated works. The study argues that the Copyright Act, 1957 was designed around the assumption of human creativity and therefore creates a significant legal gap when applied to AI systems that generate works independently. In particular, the author questions whether Section 2(d)(vi) — which attributes authorship to the person who “causes the work to be created” — can apply where the AI operates with little or no human direction. This is directly relevant to the present paper’s central concern about the adequacy of the Indian copyright framework in dealing with autonomous AI creation.
2. Hugenholtz, P.B. & Quintais, J.P. (2021). “Copyright and Artificial Creation: Does EU
Copyright Law Protect AI-Assisted Output?” IIC — International Review of Intellectual Property and Competition Law, 52, 1190. Hugenholtz and Quintais (2021) examine whether EU copyright law protects works created with AI assistance. The authors conclude that where a human author exercises sufficient creative input, the resulting work can attract copyright protection, but that fully autonomous AI output falls outside the scope of existing frameworks. This comparative perspective is useful for assessing whether India’s approach under the Copyright Act, 1957 is consistent with or divergent from international practice.
3. Pokhriyal, A. & Gupta, V. (2020). “Artificial Intelligence Generated Works under
Copyright Law.” NLUJ Law Review, 6(2). This article represents some of the earliest Indian academic writing on the subject of how copyright law should treat works of AI creation. Pokhriyal and Gupta assert that India’s Copyright Act of 1957 is inadequate in addressing the issue of autonomous AI creation. The authors assert that legislative intervention is needed to create a new category of copyright for
works of AI creation, citing the UK’s Copyright, Designs and Patents Act of 1988 as precedent. The British Act grants authorship of computer-generated works to the individual who takes steps to bring the work into being, which mirrors the goal of the current paper to discuss reforms to Indian copyright law. Taken together, these three works identify the same core legislative gap from different angles — Naithani from an Indian doctrinal perspective, Hugenholtz and Quintais from a comparative EU standpoint, and Pokhriyal and Gupta through a reform-oriented lens — and collectively support the argument that legislative intervention in India is both necessary and overdue. Case Laws
1. Eastern Book Company & Ors. v. D.B. Modak & Anr., (2008) 1 SCC 1 (Supreme Court of
India) This Supreme Court decision is central to any discussion of originality under Indian copyright law. The Court rejected the purely labour-based “sweat of the brow” doctrine and adopted an intermediate standard requiring the author to demonstrate “skill and judgment” along with a modicum of creativity. This standard directly bears on the AI authorship debate: if the creativity in an AI-generated work cannot be traced to a human author’s skill and judgment, the work may fall outside the scope of copyright protection entirely. The case thus sets the originality benchmark against which AI-generated works must be assessed under Indian law.3
2. Thaler v. Perlmutter, No. 22-cv-01564 (D.D.C. 2023) (United States District Court)
In this US decision, the District Court upheld the Copyright Office’s refusal to register an artwork autonomously created by an AI system called DABUS, holding that copyright requires human authorship. The Court stated that copyright law has never extended to works produced by non-human actors. For India, this decision is a useful comparative reference when considering the limits of Section 2(d)(vi) of the Copyright Act, 1957.
RESEARCH OBJECTIVES
1. To understand the meaning of ‘author’ under Copyright Act, 1957.
2. To explore and examine how does AI work.
3. To examine the legal framework for copyright protection in work created by AI.
3 Eastern Book Co. v. D.B. Modak, (2008) 1 SCC 1 (India).
4. To address the challenges associated with determination of authorship and ownership
rights in AI generated work.
RESEARCH QUESTIONS
1. What is the meaning of ‘author’ under Copyright Act, 1957?
2. How does AI work?
3. What is the legal framework for copyright protection in work created by AI?
4. What are the challenges associated with determination of authorship and ownership
rights in AI generated work?
RESEARCH METHODOLOGY
Doctrinal method has been used in this research paper, focusing on the critical analysis of legal principles. This method is appropriate given that the central questions are ones of statutory interpretation and identification of legislative gaps, rather than empirical inquiry. Both primary and secondary sources of data have been used. Primary sources include the Copyright Act, 1957, relevant case laws such as Eastern Book Company v. D.B. Modak and ANI Media v. OpenAI, and parliamentary reports. Secondary sources include journal articles and academic commentaries. A comparative approach has also been adopted where necessary, drawing on legal developments in the United Kingdom, United States, and European Union to inform the recommendations made in this paper.
DEFINITION OF “AUTHOR” UNDER THE COPYRIGHT ACT AND HOW DOES AI
WORK?
Section 2(d)(vi) states, "for any literary, dramatic, musical, or artistic work that is computergenerated, the person who causes the work to be created" shall be considered as an "author". Again, the copyright claim is limited by the statutory meaning of "personhood" which means the “authorship” should be given to living persons only. The independent nature of the generative AI means that it works on its own without the direct intervention of the human. This leads to the possibility that a single work can be attributed to more than one author. It has been argued that the author of the work could be the programmer who designed the AI algorithm and the person providing the prompt. Co-authorship of the work could be inferred if the programmer designed the algorithm in such a way that the user’s instructions have minimal influence on the algorithm.
Generative AI has moved well beyond the role of a simple processing tool. Unlike earlier
software that merely organised or reproduced existing data, modern AI systems are capable of producing entirely new creative works be it literary text, visual art or music, in response to a user’s prompt, without any step-by-step human direction during the creative process itself. The output of such systems, if produced by a human author, would ordinarily qualify for copyright protection on grounds of originality. This shift in the nature of AI-generated output is what makes the legal question so pressing: the law did not anticipate a machine that could generate protectable creative expression, and the Indian Copyright Framework, which has historically treated human authorship as a given, must now confront that reality directly.
INDIA’S
LEGAL
FRAMEWORK
FOR
COPYRIGHT
PROTECTION
IN
REFERENCE TO THE WORK CREATED BY ARTIFICIAL INTELLIGENCE
The existing legal regime in India has not clearly addressed the issue of copyright protection of works created by artificial intelligence. The Copyright Act of 1957 has clearly associated authorship with a human being who has created an original work. The use of artificial intelligence has been associated with automated processes. Therefore, it has not clearly recognized artificial intelligence as a legal author of a work. The use of artificial intelligence has caught the attention of policymakers in India. The 161st Report of the Parliamentary Standing Committee on Commerce has highlighted the importance of intellectual property rights in the context of emerging technologies such as artificial intelligence.
The report has recommended that the government should consider examining the existing intellectual property rights regime to ensure that it has the ability to address the legal issues arising out of artificial intelligence. 4 On the one hand, the Government of India has clearly indicated its intention to regulate artificial intelligence to ensure digital users’ rights. The statements of the policymakers, including the Minister of State for Electronics and Information Technology, have highlighted the need to develop regulatory mechanisms to ensure that innovation and accountability go hand in hand. The copyright of works generated by artificial intelligence remains uncertain under the existing Indian regime. Another important issue that arises in the context of artificial intelligence is determining the ownership of works generated by artificial intelligence.
The ownership of a work under copyright has traditionally been associated with the author of the work. The author of a work 4 Parliamentary Standing Committee on Commerce, 161st Report on Intellectual Property Rights Regime in India 42 (2021).
generated by artificial intelligence may include various parties involved in the creation of the work. These parties may include individuals who have developed the software used to create artificial intelligence. The ownership of a work generated by artificial intelligence would be a complex issue. One analogy that is frequently employed in this context is that of artificial intelligence systems and cameras in photography. Under this analogy, it can be stated that the developer of an artificial intelligence system is similar to a manufacturer of a camera, whereas the user of the artificial intelligence system who provides certain instructions is similar to a photographer who takes a picture. Under this context, it can be stated that the user of an artificial intelligence system who provides certain instructions to it can be considered comparable to an author.
Nevertheless, it can also be stated that as certain creative decisions are involved in developing an artificial intelligence system, the developer of an artificial intelligence system is comparable to an author, as opposed to a manufacturer of a camera. It can be stated that there is a great degree of uncertainty in the Indian copyright laws in light of the development of artificial intelligence. Even though it is acknowledged under the existing laws that computer-generated works can be protected, there is a need for certain legal guidelines to be framed in this context.
POSSIBILITY TO GET COPYRIGHT PROTECTION FOR AI GENERATED
CONTENT
The debate over whether AI-generated work is eligible for copyright protection is one of the most pressing issues in contemporary IP law. The concept of copyright is traditionally linked to the creator, who is the original owner of the work. The Copyright Act of 1957 did not address the issue of AI-generated work or acknowledge the concept of AI as the creator of the work. The concept of originality is an essential part of the concept of copyright. Section 13 of the Copyright Act of 1957 states that there is copyright in “original literary, dramatic, musical, and artistic works. ” Although the Copyright Act of 1957 did not define the concept of originality, the Indian courts have held that there must be at least some degree of creativity and intellectual input from the author of the work.
The problem that arises here is whether the creativity can be attributed to the author when the work is created using AI. The issue becomes complicated when the provision regarding computer-generated work is examined. The Copyright Act of 1957 was amended in 1994, and the provision regarding computer-generated work introduced the concept of the author being the person who “causes
the work to be created. ” The provision was introduced to acknowledge the role of computers in the creation of work; however, the provision assumes that there is at least some human intervention in the creation of the work. Therefore, as modern generative AI technology continues to improve, the old concepts of ownership are being turned on their head. This is because these modern AI technologies are being trained on massive datasets. This enables them to create texts, images, music, and other forms of creative works with little human intervention. This makes it unclear what exactly constitutes causality in terms of creation. Many people play a part in the creation of an AI work. This includes the programmers that create the code, the organizations that provide the training data, and the end user that creates the work or gives instructions.
In many cases, these image generators and language models are trained on massive libraries comprising millions of existing works. This brings into question concepts of originality and copyright, especially if the works on which the AI is trained include copyrighted materials — even if the works that the AI ultimately creates are themselves original. There is also the question of what constitutes the definition of "person" in modern copyright law. This is because, currently, only natural people or organizations can own copyrights. This is in line with the fact that AI systems do not have legal personalities. This makes it such that if an AI is defined as the author, it is unclear what exactly that would mean. The concept of ownership is further complicated by the fact that many people play a part in the creation of an AI work.
This includes the programmers that create the code, the organizations that provide the training data, and the end user that creates the work or gives instructions. This makes it unclear what exactly constitutes authorship under existing copyright laws. One notable instance from India highlights a major issue that can arise when it comes to AI authorship. This case revolves around an AI tool called RAGHAV. In 2020, the copyright registration process was successfully completed for a piece of artwork that was created with the help of RAGHAV. In the copyright claim form, the author had mentioned that the artwork was created with the help of the AI tool RAGHAV and that the author himself was the creator along with the AI tool. Later on, the Indian Copyright Office sent a notice to the author stating that the AI tool could not be considered the author of the artwork.
According to them, the copyright law only recognizes the authorship of a natural person or a legal entity. Therefore, although the Copyright Act has provided limited protection to computer-generated works, it still does not address the complexities of autonomous works.
The situation in India has its parallel in the United States. In the US, in the case Thaler v. Perlmutter (No. 22-cv-01564, D.D.C. 2023), the US District Court upheld the Copyright Office's decision not to grant registration for a work entirely created by the AI system DABUS. The US court reaffirmed the well-established rule that the grant of copyright depends on human authorship. This rule is well entrenched in the US Constitution and Copyright Act. What this means for the Indian courts is that if the work is created entirely by the AI system with no human creative input, there is no human author and therefore no copyright protection, irrespective of the aesthetic appeal of the work. This decision also confirms the position that the Copyright Act, 1957, in its present form does not grant protection for works created entirely by autonomous AI systems.5
CHALLENGES IN DETERMINING AUTHORSHIP AND OWNERSHIP RIGHTS
The speed at which artificial intelligence is being developed has raised legitimate concerns over who owns the product of AI. Currently, the copyright law protects the programming code and the databases used in developing AI. However, if the law grants the developer of the AI the exclusive rights over the copyright of all the products generated by the AI, then the law is likely to create too many copyrights. AI has the capability of producing numerous products, and granting the developer exclusive rights over all the products generated by the AI may be beyond the capability of the copyright law, which is supposed to encourage creativity and the free flow of knowledge among the public. Another issue concerns the relationship between AI developers and the users of their systems.
Many large AI platforms, including OpenAI, have adopted terms of service under which the developer does not claim ownership of the output generated by users. However, these same terms acknowledge that the system may produce identical or substantially similar outputs for different users entering the same prompt. This means that no single user can claim exclusive ownership over the output, since another person may have independently received the same result. The copyright framework currently has no mechanism to address this form of nonexclusive, simultaneously generated output, which sits uneasily with the principle that copyright protects a unique creative expression. Moreover, the development of large language models and the generative AI has led to copyright disputes in many parts of the world.
The rise of AI has made many creators worried that their work is being used in the development of AI models without permission. The development of
5 Thaler v. Perlmutter, No. 22-cv-01564, slip op. at 8 (D.D.C. Aug. 18, 2023).
AI models requires huge databases of materials, including books and artworks, which are subject to copyright law. Several authors have expressed such apprehensions. First, the use of copyrighted work in the training data set without permission could constitute the unauthorized use of protected work. Second, the output generated by AI could mimic or copy certain features from other works, thus infringing on the moral rights of the original creators. Third, the proliferation of AIgenerated work could threaten the livelihoods of writers, artists, and other creative individuals, who could lose work that would have required human creativity. Fourth, the authors argue that if their work is being used for training AI, they are entitled to some form of compensation. These concerns are no longer merely theoretical. In Andersen v. Stability AI Ltd. (No. 3:23-cv- 00201, N. D. Cal.
2023), a class of visual artists in the United States filed suit against multiple AI image-generation companies, alleging that their copyrighted works had been scraped from the internet and used without consent to train generative AI models. The plaintiffs contended that this constituted direct copyright infringement, and that the AI outputs themselves were unauthorised derivative works. The case has significant comparative relevance for India: it illustrates that the question of training data liability is a live legal issue worldwide and one that the Copyright Act, 1957 is presently ill-equipped to resolve, given the absence of a specific text-and-data-mining exception or AI-specific fair dealing provision. 6 Closer to home, India has also witnessed its first major copyright dispute involving AI and training data. In ANI Media Pvt. Ltd. v. OpenAI Inc.
(CS(OS) 3766/2024, Delhi High Court), the news agency ANI filed a suit against OpenAI, alleging that its copyrighted news articles had been used without authorisation to train the ChatGPT large language model. The Delhi High Court, while hearing the matter, raised important questions about the territorial application of Indian copyright law to AI training activities conducted abroad and about the adequacy of the existing statutory framework. This case represents a landmark moment for Indian copyright law and underscores the urgent need for legislative clarity on the permissibility of using copyrighted works in AI training datasets. 7 The second set of issues relates to whether artificial intelligence can be recognized as an author at all. If such recognition is made, there are substantial legal implications that could arise.
The first is that AI lacks legal personality; therefore, it cannot hold any form of property, contract,
6 Andersen v. Stability AI Ltd., No. 3:23-cv-00201 (N.D. Cal. Oct. 30, 2023).
7 ANI Media Pvt. Ltd. v. OpenAI Inc., CS(OS) 3766/2024 (Delhi H.C. 2024).
or liability. If there is infringement of copyright through the creation of AI-generated work, who do you hold liable if the AI is recognized as the author? Additionally, there are Indian law complications that are specific to the Indian jurisdiction. Literary, dramatic, musical, and artistic works are protected for sixty years after the death of the author. If the AI is recognized as the author, determining the exact period of protection would be difficult, considering that the AI has no lifespan. Apart from the issue of copyright infringement, the use of generative AI has raised concerns about personality rights and the unauthorized use of a person’s identity. One such case in India is the case of Anil Kapoor v. Simply Life India & Ors. , in which the actor Anil Kapoor has filed a case in the Delhi High Court seeking protection for his personality rights.
Kapoor claimed that his name, image, voice, and other aspects of his personality, such as his catchphrases and nicknames such as “Lakhan,” have been used by online websites and AI technology without his authorization. The Delhi High Court has issued an interim injunction against the unauthorized use of the actor’s name, image, voice, and other aspects of his personality. The Court has held that the use of technology, especially AI, can be used to infringe on a person’s personality rights. Although the use of AI is for lawful purposes, the commercial exploitation of a person’s personality through AI-generated content without authorization can be considered a violation of personality and publicity rights. The above case has demonstrated that the emergence of AI has raised complex legal issues that go beyond the issue of copyright infringement.
Such issues of authorship, ownership, and the protection of personality rights have remained unanswered in the existing legal scenario. It has, therefore, become imperative that proper legal guidelines be framed in this regard. A subsequent case further consolidated the recognition of personality rights as a distinct legal
concern in the AI era. In Jaikishan Kakubhai Saraf (Jackie Shroff) v. The Peppy Store & Ors. (CS(OS) 1016/2024, Delhi High Court), the actor Jackie Shroff sought protection against the unauthorised use of his name, image, voice, and distinctive style of speech by AI chatbots and other digital platforms. The Delhi High Court granted an ex parte injunction restraining the defendants from using the plaintiff’s persona in any AI-generated content without his consent. Taken together with the Anil Kapoor decision, this case signals the emergence of a judicial recognition in India that personality rights — encompassing name, likeness, and voice — must be protected against AI-enabled misappropriation, even in the absence of a dedicated statutory
framework. These cases make a compelling argument for the legislative recognition of personality rights within the broader framework of intellectual property law in India.8
CONCLUSIONS AND SUGGESTIONS
The growing capabilities of generative artificial intelligence have exposed a deep tension at the heart of Indian copyright law. A legal framework built around human creativity is being asked to govern a world in which machines can independently produce literature, art, and music. This paper has examined that tension across four dimensions: the definition of authorship under the Copyright Act, 1957; the existing legal framework for copyright protection of AI-generated works; the challenges in determining authorship and ownership; and the emerging issues of training data liability and personality rights. Three main conclusions emerge from this analysis. First, the Copyright Act, 1957, in its current form, cannot accommodate autonomous AI-generated works. Section 2(d)(vi) was drafted at a time when computers were tools that operated entirely under human direction.
Modern generative AI, which produces creative output from a single prompt without step-by-step human involvement, does not fit that model. The causal link between human instruction and creative output that the provision presupposes has broken down. Furthermore, the Supreme Court’s requirement of “skill and judgment” as articulated in Eastern Book Company v. D. B. Modak is difficult to satisfy where the creative choices are made by the AI system rather than by any identifiable human author. Second, the question of ownership of AI-generated works involves several stakeholders - the developer of the AI system, the organisation providing the training data, and the end user providing the prompt, none of whom clearly fits the definition of “author” under existing law.
The camera analogy, which compares the AI developer to a camera manufacturer and the user to a photographer, offers a useful starting point but ultimately falls short. A camera does not make independent creative decisions; a generative AI system does, and the law has not yet found a satisfactory way to characterise that distinction. Third, the challenges extend well beyond authorship. The use of copyrighted works to train AI models, as seen in ANI Media Pvt. Ltd. v. OpenAI Inc. in India and Andersen v. Stability AI Ltd. in the United States raises serious questions about training data liability that the Copyright Act, 1957 is not equipped to answer. Furthermore, the Delhi High Court’s recognition of 8 Anil Kapoor v. Simply Life India & Ors. , CS(OS) 3278/2023 (Delhi H. C. 2023); Jaikishan Kakubhai Saraf v. The Peppy Store & Ors. , CS(OS) 1016/2024 (Delhi H. C. 2024).
personality rights in Anil Kapoor v. Simply Life India & Ors. and Jaikishan Kakubhai Saraf v. The Peppy Store & Ors. confirms that the legal challenges posed by generative AI reach beyond intellectual property into questions of individual dignity and identity. Suggestions In light of the foregoing, this paper makes the following five legislative and policy recommendations for reforming Indian copyright law to address the challenges of artificial intelligence. Recommendation 1: Amendment to Section 2(d)(vi) — Tiered Authorship Framework. The Copyright Act of 1957 needs to be amended to provide a distinction between AI-assisted and AI-generated works.
In cases where a human author has control over the final product, for example, where they make certain artistic choices in the creation of the work, as opposed to simply entering a prompt for the AI to fulfil, they should have full authorship and copyright of the work as it is defined today. In cases where the work is more autonomous in its creation, the copyright should rest with the person who has made the necessary arrangements for the creation of the work. This is based on the "arrangements" standard set in Section 9(3) of the UK Copyright, Designs and Patents Act of 1988, which has already been implemented for computer-generated works. This will help to eliminate the confusion in Section 2(d)(vi) without necessarily granting legal person status to the AI system itself. 9 Recommendation 2: Reduced Term of Protection for AI-Generated Works.
Where the copyright in the work generated by the AI vests in a legal entity under the proposed system, the duration of the protection must be significantly lower than the standard sixty-year period applicable to human authors. A fixed period of fifteen to twenty-five years from the date of first publication is suggested. This is based on two considerations. Firstly, the rate and volume at which works can be generated by an AI far surpass any human author’s capacity, and the duration of protection would lead to an unprecedented level of concentration of copyrights in the hands of the authors. Secondly, the purpose of copyrights as a means to return creative works to the public domain over time requires that works generated by AI do not dominate the public domain for centuries to come.
Recommendation 3: Mandatory AI Disclosure and a Registration Scheme Any person seeking copyright protection for an AI-generated work should be required to disclose, at the time of registration with the Copyright Office, that the work was produced using an AI system. The disclosure should specify the AI system used and the degree of human
9 Copyright, Designs and Patents Act 1988, § 9(3) (UK).
creative involvement in the process. The Copyright Office should maintain a separate register for AI-generated works, distinct from the existing register for human-authored works. This serves three practical purposes: it promotes transparency and deters fraudulent claims of human authorship; it assists courts in resolving ownership and infringement disputes; and it provides policymakers with reliable data to assess and refine the legislative framework over time. Recommendation 4: Introduction of a Text and Data Mining Exception with a Remuneration Mechanism. Parliament should introduce a text-and-data-mining exception to the Copyright Act, 1957, permitting AI developers to use lawfully accessed copyrighted works for the purpose of training AI models.
Without such an exception, developers face genuine legal uncertainty, as reflected in the pending Delhi High Court proceedings in ANI Media Pvt. Ltd. v. OpenAI Inc. However, this exception must be paired with a statutory remuneration mechanism, administered through a collecting society, to ensure that rights holders receive fair compensation when their works are used for commercial AI training. This approach draws from Articles 3 and 4 of the EU’s Digital Single Market Directive, 2019, and would address the legitimate economic concerns of Indian authors, journalists, and artists whose works are routinely used in AI training without their knowledge or consent. 10 Recommendation 5: Statutory Recognition of Personality Rights. In this regard, the Delhi High Court has recognized personality rights through judicial innovation in Anil Kapoor v. Simply Life India & Ors.
, and Jaikishan Kakubhai Saraf v. The Peppy Store & Ors. , restraining the unauthorized use of a person's name, image, and voice in AI-generated content. Nevertheless, judge-made laws are inherently uncertain and inaccessible to ordinary individuals, who cannot afford litigation costs. Therefore, Parliament must enact a separate legislative provision, either in the Copyright Act, 1957, or in a separate legislation, specifically prohibiting the use of any person's identity in AI-generated content in the absence of their informed consent. Further, the legislation must specifically include deepfakes, voice cloning, and any other technology that can reproduce a person's identity. Such legislation will not only safeguard the rights of celebrities, but also offer effective protection to every citizen from AI-generated identity theft.
10 Council Directive 2019/790, arts. 3–4, 2019 O.J. (L 130) 92 (EU).
BIBLIOGRAPHY
Andersen v. Stability AI Ltd. , No. 3:23-cv-00201 (N. D. Cal. 2023). • ANI Media Pvt. Ltd. v. OpenAI Inc. , CS(OS) 3766/2024 (Delhi High Court). • Anil Kapoor v. Simply Life India & Ors. , CS(OS) 3278/2023 (Delhi High Court). • Copyright Act, 1957 (India), as amended by the Copyright (Amendment) Act, 1994. • Copyright, Designs and Patents Act, 1988 (United Kingdom), s. 9(3). • Eastern Book Company & Ors. v. D. B. Modak & Anr. , (2008) 1 SCC 1 (Supreme Court of India). • European Parliament and Council, Directive (EU) 2019/790 on Copyright in the Digital Single Market (DSM Directive), Articles 3–4. • Hugenholtz, P. B. & Quintais, J. P. (2021). “Copyright and Artificial Creation: Does EU Copyright Law Protect AI-Assisted Output? ” IIC — International Review of Intellectual Property and Competition Law, 52, 1190–1216. • Jaikishan Kakubhai Saraf (Jackie Shroff) v. The Peppy Store & Ors.
, CS(OS) 1016/2024 (Delhi High Court). • Naithani, P. (2022). “Issues of Authorship and Ownership in Work Created by Artificial Intelligence — Indian Copyright Law Perspective. ” NTUT Journal of Intellectual Property Law & Management, 11(1), 7–28. • Parliamentary Standing Committee on Commerce, 161st Report on Intellectual Property Rights Regime in India (2021), Rajya Sabha. • Pokhriyal, A. & Gupta, V. (2020). “Artificial Intelligence Generated Works under Copyright Law. ” NLUJ Law Review, 6(2), 51–74. • Thaler v. Perlmutter, No. 22-cv-01564 (D. D. C. Aug. 18, 2023).