AI And Copyright In India - Rethinking Authorship In The Digital Age
Author: LawSphereX
Category: Technology, Cyber & Privacy
Published:
Updated:
AI AND COPYRIGHT IN INDIA: RETHINKING
AUTHORSHIP IN THE DIGITAL AGE
Pari Gupta, Vivekananda Institute of Professional Studies “Creativity is part of human nature. It can only be untaught.” - Ai Weiwei'
ABSTRACT
Human creativity, with its inherent originality, emotional depth, and personal touch has historically been the basis of copyright law. However, now that generative Artificial Intelligence (AI) with an ability to generate literature, music, and visual art based on minimal human prompts, AI is forcing us to rethink our assumptions about authorship. AI and its products challenge the essential principle stated above because copyright is meant to acknowledge and protect human creative work. In India, the Copyright Act of 1957 provides protection for computergenerated works, as outlined in Section 2(d)(vi). This section attributes the authorship to "the person who causes the work to be created”. Although this was quite forward-thinking at the time it was introduced, it was conceived during an era when automation was still in its infancy.
This article employs a doctrinal research methodology and qualitative content analysis to examine laws, court cases, and policy discussions in India. Keywords: AI-generated works; Copyright Act 1957; Computer-generated works; Significant Human Input; Training Data; Indian Copyright Law.
1. INTRODUCTION
The concept of creativity is always considered as one of the fundamental approaches to the definition of what it means to be human. Consider a cave painting of a prehistoric period or the newest blockbuster movie, both have imagination, feeling, and purpose and these are aspects that are obviously not part of mechanical process outputs. Due to this, copyright law has been enacted to protect such works and enable the authors to make money out of their ideas and culture. Generative AI is now disrupting things. Such tools as ChatGPT, MidJourney, Gemini and other models can generate works that virtually resemble human writing and work. Here, the distinction is that unlike the older assistive technology, they are to a large extent capable of working on their own, with only a briefing. And then that immediate leap raises questions: Is AI-written content subject to copyright?
And being so, who is the owner?
2. THE INDIAN LEGAL FRAMEWORK
2.1 Authorship under the Indian Copyright Act, 1957
The centre of the authorship debate is the Section 2(d) that describes the definition of author in relation to different genres of works: • In literary or dramatic compositions: the individual who is the source of their production; • In the case of works of art: the individual who designs the work; • In the case of a computer-generated literary, dramatic, musical or artistic work: the one who instigates the creation of the work (Section 2(d)(vi)). Amendments made in 1994 have inserted section 2(d)(vi) which states that where the work is computer-generated, “the author” is that person who causes the work to be created1. It seems like a fairly progressive definition, even though this definition was published in 1994, which implies that policymakers had already become aware of the concept of non-human creativity, long before the current generative AI debate.
And the law has never defined what is meant by
1 The Copyright Act, 1957, No. 14, Acts of Parliament, 1957 (India), Section 2(d)(vi)
causing in the context of more advanced AI where both human and non-human actors are involved. Section 17 presupposes that it should not be difficult to determine the identity of the author; that the author is most likely the original owner of copyright, except in a situation where the work has been produced in a contract of employment or a commissioned piece of work in which case it transfers elsewhere. That assumption of sense is currently shaken by works where AI is either a collaborator or an aide to human artists.
2.2 Judicial Interpretation on Originality
In Indian copyright law, there has been a significant shift from an effort-based criterion to one that requires a measurable degree of creativity. This evolution is crucial especially when considering how the law can protect works generated by Artificial Intelligence. The Privy Council laid out the basic standard of the doctrine in Macmillan & Co. v. K. & J. Cooper 1, (1924)2, which involved the doctrine of sweat of the brow. The Court has warned that originality is not always that of novelty; it requires, instead, skill, labour, and judgment. The case concerned educational works based on an existing material. Quite the contrary, the Court recognized that, when existing material is otherwise rendered, the selection, coordination, and presentation involved therein, when carried out with the exercise of some human judgment, may pass on to the resulting work originality.
This early stance made industriousness its pinnacle of merit and not innovative creativity. The Macmillan test became predominant in Indian copyright law, until being abrogated by the Supreme Court of India in the court case of Eastern Book Company v. D. B. Modak in 20083. In Eastern Book Company, the Court departed the radically “sweat of the brow” test and developed a “modicum of creativity” approach, citing an American classic, Feist Publications Inc. v. Rural Telephone Service Co4. Unlike the requirement of originality in the case of the U. S, where inventive genius is required in order to satisfy originality, the Indian doctrine presently solely requires that the work should possess (not much more than) a minimum of intellectual endeavour beyond organisation or labour of a skill-based nature.
2 Macmillan & Co. v. K. & J. Cooper, (1924) 26 BOMLR 292, 296 (PC)
3 E. Book Co. v. D.B. Modak, (2008) 1 SCC 1
4 Feist Publ’ns, Inc. v. Rural Tel. Serv. Co., 499 U.S. 340, 345–46 (1991)
The decision of the Court to consider human intellectual contribution further comes out in the cases of Navigators Logistics Ltd. v. Kashif Qureshi & Ors5. In this case, a computer generated list of customers was reviewed and the Court held that even though the list was prepared using computer, it did not amount to originality as required under the Copyright Act, 1957 since no significant human input had gone into its creation. Even though these compilations can be vast with commercial benefit, they will not be guarded when there is no indication that human ingenuity has designed the end product. This principle becomes more important during the modern age of the artistic presence of AI, where a human can become involved in the creation of art visibly little (only by providing a prompting) or in an editor capacity (by editorial curation)6.
A combination of these rulings delineates that in the Indian law originality lies in the
intellectual input of humans. Via placing the emphasis on the originality in skill and labour (Macmillan), a higher score on minimum creative expression (Eastern Book Company) and the omission of works wholly created through mechanical hands (Navigators Logistics) these fairly illustrate the judicial course of originality in India. These precedents will guide the future judicial thought process and challenge the resilience of a human-friendly system of copyright in a more technologically advancing creative industry7.
3. WHO QUALIFIES AS THE AUTHOR OF AI GENERATED WORKS?
The major challenge in AI based works lies not in the question of whether AI based works can be copyrighted or not, but also in the question of who should be the recognised author to the work. Section 2(d)(vi) identifies the act that causes the creation of a computer-generated work, the person doing so will be considered the author, usually understood as the person or group that gives the instructional input or the person or group that develops the system. But now that AI generates a complex output, such as a painting or a novel, with little or no input, that longstanding focus on human agency is being challenged8. Take an example of an AI-driven music application where a customer commands the
5 Navigators Logistics Ltd. v. Kashif Qureshi & Ors., (2014) SCC Online Del 4183
6 Harshal Chhabra & Kanishk Gaurav Pandey, Balancing Indian Copyright Law with AI-Generated Content:
The ‘Significant Human Input’ Approach, IJLT (Feb. 26, 2024) 7 Shubham Sharma, Rethinking Originality in the Age of AI: Challenges for Indian Copyright Law, 5 INDIAN J.L. & TECH. 123, 127–30 (2023).
8 AI as Creative Collaborator: Rethinking Authorship and Copyright in the Digital Era, INT’L J. INNOVATIVE RES. L., Jan. 2025, at 45–47.
application with a basic request like, compose me a classical raga with a touch of modern jazz. The system assembles a fully orchestrated composition, but with a combination of Hindustani melodic structures and jazz harmonies, without any human guidance visible. By a literal approach of interpreting Section 2(d)(vi), the user of the system would be considered the author. However this interpretation marginalizes two other factors: the programmers who created the structure of the AI model, and the sheer number of musicians whose recorded music was used as training data. Such situations point to the insufficiency of the law and legal doctrine that assumes human input and creative output has a simple cause and effect relationship. AI disputes this claim since its outputs usually come out of opaque decision-making processes, what is often termed as the black box problem.
Generative AI work is not easily traceable to particular human actions of creativity since unlike previous software, the output is produced by a computer and can be ultimately attributed to one or more human actions. Currently, the authorship of AI-generated content is being assigned to the agency of human subjects - most likely the users or creators - rather than to the AI systems themselves. This solution represents the general view in the world that machines cannot have the legal personality. That said, with generative AI systems perhaps previously lowering the degree of human supervision to render creative work, India might soon encounter the issue of reevaluating and perhaps revising Section 2(d)(vi) to determine whether the user, the programmer, or no one in particular need not be considered the legal author of the creatively produced work9.
4. EMERGING LEGAL RESPONSES TO AI-GENERATED WORKS
Recent events in India highlight the changing concerns of the application of copyright law to AI-generated content and training materials. Two interesting cases, the RAGHAV case and ANI v. OpenAI, mark the ambiguity in authorship and infringement in the AI age. The case of RAGHAV concerned an artificial intelligence tool called Robust Artificially Intelligible Graphics and Art Visualizer (RAGHAV), which creates art. In one instance, a copyright registration application was filed naming RAGHAV as the exclusive author which 9 Vidhi Shah & Aishvi Shah, Artificial Intelligence and Copyright Law: An Entreaty for Legal Framework with Special Reference to TRIPS Agreement, 1 IPR J. MAH. NAT’L L. U. NAGPUR 33, 41–42 (2023)
was refused by the Indian Copyright Office with reference to the necessity of a human author by the provisions of Section 2(d)(vi). In another instance, human was designated as co-author in RAGHAV and although the registration had been initially allowed, it was revoked. These formal measures can be seen as the Copyright Office being reluctant to acknowledge AI as an author or co-author, the uncertainty of applying the current legislation to AI-generated content and the necessity to provide more specific guidance on the authorship of a written work that was created jointly by a human and a machine10. In January 2024, OpenAI case (CS(COMM) 228 /2024) became a landmark in AI-copyright jurisprudence in India. ANI, a prominent news agency, accuses OpenAI of using its copyrighted newspaper articles to educate ChatGPT and the software can replicate ANI newspaper material without permission.
ANI invokes Sections 14 and 51 of the Copyright Act, in order to seek injunctions, damages, and disclosure of training data. OpenAI justifies its activities as reasonable dealing under Section 52(1)a(a) of research and as one of transformative use, comparable to human education. In March 2025, Delhi High Court granted an interim injunction against OpenAI, claiming an infringement of ANI content by prima facie on the fact that no express TDM exemptions were present. The legal aspects of data scraping to train AI and the limits of fair dealing are investigated in this case as the hearings are continued until as late as September 2025. A decision in favour of ANI may make it obligatory to license training data, enhancing the rights of creators but potentially restricting the availability of AI in India11. Combined, the ANI v.
OpenAI and RAGHAV cases highlight the urgency of the legislative change in the understanding of how AI-produced works are treated, who should get the credit of their creation, and whether it is even legal to train AI using a piece of copyrighted writing. In the absence of statutory action, courts and regulators will persist in addressing each case separately, and creators, AI developers and the publishing industry will all live in uncertainty12.
5. CHALLENGES OF AI-GENERATED WORKS IN INDIAN COPYRIGHT LAW
Generative AI introduces new issues to the copyright system in India, specifically how to define authorship and overcome the possibility of AI training data causing infringement. These
10 Dhruv Verma, Mind Over Machine: Reimagining Indian IP Law in the Age of AI, JURISTS’ JUNCTION
(Aug. 22, 2025)
11 Asian News Int’l v. OpenAI, CS(COMM) 228/2024 (Delhi High Ct. 2024) (India)
12 Harshal Chhabra & Kanishk Gaurav Pandey, Balancing Indian Copyright Law with AI-Generated Content: The ‘Significant Human Input’ Approach, INDIAN J.L. & TECH. (Feb. 26, 2024)
advancements work against the scope of the Copyright Act, 1957 which requires adequate examination and potential review of its ideologies13.
5.1 Authorship and Originality
Tools of generative AI, be they of literature, music, or visual art, often require minimal human oversight, which makes determining authorship under Section 2(d)(vi) challenging. As an example, a user prompt such as write a Bollywood-style screen play can result in a script that is written in detail solely due to the algorithms of the AI. To determine whether the input provided by the user can meet the requirement of the modicum of creativity, as would be required in Eastern Book Company v. D. B. Modak (2008)14, could be a challenge on the side of the courts. In case human interference is considered to be inadequate, there is a risk that the product would fall outside of the copyright protection and saturate the common sphere, which would discourage the human abilities to produce anything.
On the other hand, making users the full authors when they have only prompted a little might be rewarding them excessively and leave out programmers or creators of the underlying programs.
5.2 Training Data and Infringement Risks
The second problem is the difficulty in training AI models using copyrighted content. The process of generative AI systems is based on large volumes of data, in most cases, collected by crawling the internet without the consent of the author, which is problematic in connection with the provision of Section 51 of the Act, which determines infringement. In contrast to other jurisdictions such as the European Union that introduced text-and-data mining (TDM) exemptions in Directive 2019/790, Indian practices, through Section 52 of Copyright Act, do not provide commercial-AI-training and therefore leaves developers susceptible to liability. The current case of Asian News International (ANI) v. OpenAI15 highlights this tension. ANI claimed that its articles were used to train ChatGPT without the company's permission.
This case brings up the necessity of a definitive ruling on whether AI training falls under fair dealing in sub-clause 52(1) (a) where the research purposes are involved or it is an infringement. 13 Nikhil Mishra & Digvijay Singh, AI-Generated Work and its Implications on Copyright Law in India, 30 J.
INTELL. PROP. RTS. 35, 36 (2025)
14 E. Book Co. v. D.B. Modak, (2008) 1 SCC 1
15 Asian News Int’l v. OpenAI, CS(COMM) 228/2024 (Delhi High Ct. 2024) (India) (ongoing as of Sept. 2025).
6. FINDING AN INDIAN MIDDLE PATH: POSSIBLE MODELS FOR REFORMS
The copyright Act of 1957 in India has the problem of adapting AI-generated works to the law that has traditionally focused on human authors. This will need a more subdued, middle-ground approach, an approach that acknowledges the value of AI in creativity but safeguards human creators and makes it available to the broader population16.
6.1 Collaborative Authorship
The unification of the authorship between the users and programmers could be formalized by amending Section 2(d) and introducing the concept of shared authorship between the users and the programmers by acknowledging the existence of a human-AI consortia. Human authors who made significant contributions to the work (through elaborated prompts or conceptual direction) might be identified as primary authors. An AI system designer, trainer and fine-tuners may be given secondary rights, such as royalties. Metadata and the logs of algorithms can be used as evidence of contribution. The model is an expression of the Indian cultural custom of collective creativity via collective arts, where the recognition and incentives are not concentrated on the right side17.
6.2 Sui Generis Protection for Autonomous AI Works
Sui generis protection may be provided in limited term (e.g. 10-15 years) to works produced with little or no human intervention. This strategy is based on the inspiration of what India has done by protecting traditional knowledge through the Biological Diversity Act, 200218, which rewards innovations but makes outputs ultimately propagate to the public, thus preserving cultural accessibility and eliminating monopolization.
6.3 Text-and-Data Mining (TDM) Exceptions
In order to deal with copyright issues in AI training, Section 52 might be amended to permit TDM exemption to non-commercial research. In a commercial context a mandatory licensing system would probably guarantee that creators, including publishers, moviemakers and artists,
16 India’s Approach to AI Governance, NITI AAYOG, https://www.niti.gov.in/sites/default/files/2023-
02/AI_India_Governance (last visited Sept. 12, 2025). 17 AI as Creative Collaborator: Rethinking Authorship and Copyright in the Digital Era, Indian Journal of Integrated Research in Law
18 The Biological Diversity Act, 2002, No. 18, Acts of Parliament, 2002 (India)
are fairly remunerated. This moderates the innovative requirements against the protection of the creative industries which is consistent with the concerns of equity in accessing knowledge in India.
6.4 Transparency and Accountability Mandates
Disclosure of AI training datasets and algorithm processes may promote accountability and promote infringement analysis by making it easier to attribute algorithmic processes fairly. This can be enforced by a regulatory commission similar to the Geographical Indications Registry in India whereby innovation is not throttled19. This openness can find echo with the wider Indian systemic aim to uphold good governance and will foster societal confidence on AI-based technology as well as protect creative intellectual input. Through collaborative authorship, sui generis protections, TDM exceptions and accountability quotas, India can develop a middle-ground approach that acknowledges human invention, encourages technological innovation as well as sustain access to cultural knowledge.
This would place India at the forefront of the entire world in terms of ethical, inclusive, and futureoriented AI copyright protection20.
7. CONCLUSION
The drastic effect of generative AI in the realm of creativity requires a visionary revamp of the Copyright Act, 1957 in India in order to balance the high-speed innovation with the rich cultural heritage of the nation. The ANI v. OpenAI, Eastern Book Company v. D. B. Modak and Navigators Logistics Ltd. v. Kashif Qureshi and Ors. - all also underscore the necessity to balance the ethics of AI products with their usage in a very practical way through authorship, data ethics, and equity in access. This could be achieved by taking an Indian Middle Path, where India can build a copyright system that the people and the user community share rights more fully and promotes an active digital economy21.
As part of the Indian culture of balance and inclusiveness, these reforms will safeguard creators and make India a worldwide leader in 19 The Geographical Indications of Goods (Registration and Protection) Act, 1999, No. 48, Acts of Parliament,
1999 (India), Section 11.
20 Poorna Mysoor, From Creation to Collection: Navigating AI’s Role in Indian Copyright Law, 12 QUEEN
MARY J. INTELL. PROP. 89, 95–97 (2022).
21 Digital India Programme, MINISTRY OF ELECTRONICS & INFO. TECH., GOV’T OF INDIA,
https://www.digitalindia.gov.in (last visited Sept. 12, 2025).
balancing AI-driven creativity with both the law and culture, which protections will keep its legislation pace with its future technological ambitions.