When Machines Create: The Emerging Copyright Challenges of Generative AI in India

Posted On - 11 November, 2025 • By - Surbhi Kapoor

Introduction

In the twenty-first century, the applicability and advancement of Artificial Intelligence (AI) in different aspects of everyday and workplace use has exponentially increased. This has led to the birth of two main types of AI generative artificial intelligence and predictive artificial intelligence. The former one has disrupted content creation and other professional realms that are rooted in longstanding perceptions of authorship, originality, and fair use across national legal systems. Now, India’s Copyright Act of 1957 was drafted in a pre-digital society and faces unprecedented hurdles as generative models such as ChatGPT use a plethora of copyrighted content for training in order to provide rapid contextualized responses.

In May 2025, the Ministry of Commerce set-up an eight-member expert committee as the ANI Media v. OpenAI litigation continued before the Delhi High Court (marking India’s first institutional and statutory effort to incorporate post-digital copyright regimes into a pre-digital copyright regime). This article will explore how the complementary advancements reveal important gaps in the copyright law in India, specifically with regard to authorship in works created by technology and the application of the fair dealing provision in Section 52 of the Copyright Act, 1957 . This will be accomplished by exploring the expert committee’s jurisdiction.

On April 28, 2025, the Department for Promotion of Industry and Internal Trade (DPIIT) issued an Office Memorandum to establish a committee of eight members[1] to examine issues around generative technology and copyright. The expert panel consists of intellectual property lawyers, senior bureaucrats, and industry representatives – acknowledging the multidisciplinary nature of the legal uncertainties surrounding AI. Their work will touch on four topics: 1) Identifying gaps in current copyright law, 2) Examining the Copyright Act of 1957, 3) Providing policy recommendations, and 4) Creating a working document related to proposed reforms.

The formation of a panel in this context indicates the government’s recognition that the copyright system in India is in need of fundamental reassessment because the emergence and transformative use of generative AI technology for content creation is shifting the nature of the process. Unlike legislating specific provisions in the Copyright Act of 1957 governing matters around generative AI, the panel will address issues specific to generative AI, to assess whether the Copyright Act of 1957 appropriately considers: a) how access and use of data in AI training can occur without infringing copyright, b) how authorship will be attributed given works generated by AI, and c) ownership rights in works created with AI technology. In addition to these specific issues pertaining to AI, the panel will also consider broader issues regarding the applicability of fair dealing exceptions under the Copyright Act of 1957 for data extraction and preparation of data for AI operations, which are an essential component of developing large scale AI models.

The committee faced initial issues dealing with credibility when one of its members withdrew from the committee, stating the reason for their withdrawal was they did not have the appropriate technical capabilities, which indicated the challenge of bridging the legal and the technical domains. As the panel engaged in seeking solutions to questions around transparency, and representation of all stakeholders, especially content creators and developers of technology, issues arose. The timing of the committee’s first meeting on May 16, 2025, and developments in the ANI Media v. OpenAI litigation suggested a simultaneous judiciary and policy approach to the issues at hand.

Asian News International (ANI) filed a lawsuit (CS(COMM) 1028/2024) before the Delhi High Court in November 2024 that alleged the unauthorized use of its copyrighted news content to train ChatGPT without obtaining a license. ANI Media originally was seeking ₹2 crore in damages and a permanent injunction. The Court eventually framed important questions regarding whether training on or with copyrighted data infringed copyright, whether generating a response was a reproduction of copyrighted material, and whether the fair dealing provisions for large-scale data mining undertakings.

Justice Amit Bansal recognized the landmark implications of the case for copyright jurisprudence in the technological era. ANI contended that OpenAI infringed upon the Copyright Act, Sections 14 and 52, by utilizing ANI’s material for commercial purposes without consent. OpenAI’s defence presented arguments on jurisdictional grounds and focused on an assertion of compliance with global standards, further claiming that tokenization is only a process of converting text into numerical data and does not involve generating or storing retrievable copies. OpenAI asserted that its usage is transformative and not reproductive.

The case attracted significant industry participation from stakeholders such as the Federation of Indian Publishers, Digital News Publishers Association, T-Series, and Sony Music, highlighting the potential ramifications of the case upon the creative economy. Justice Bansal appointed Dr. Arul George Scaria and Adarsh Ramanujan as amicus curiae to provide assistance to the court. Their competing frameworks for permissible use illuminate tensions between copyright and innovation. The case parallels the litigation of The New York Times v. OpenAI in the United States, again illustrating the borderless nature of the case.

Fair Dealing Doctrine in India[3]

A limited form of fair dealing is found in Section 52 of the Copyright Act, 1957, which permits only restricted uses in accordance with the section’s list specifically referring to use for personal study, research, criticism, review, and reporting current events, among others. With Civic Chandran v. Ammini Amma, we can see that Indian courts have been methodical in a relatively narrow application of Section 52, as opposed to the four-factor fair use test in Section 107 of the United States Copyright Act. As transformative use is not recognized under Indian law, innovation and data processing related to AI is in a Catch-22 position.

Large-scale data mining for training AI models creates a situation where traditional boundaries of private and commercial use become challenged, where the scale of data processed makes it difficult to view “private use” exceptions for training. India does not have statutory provisions providing explicit Text and Data Mining (TDM) exceptions under common law like the EU Digital Single Market Directive. Hence, the act of training programs commercially, and systematically reproducing copies of copyrighted material for profit does not qualify as fair dealing.

Other comparative models, such as Japan’s 2018 copyright amendment and the EU TDM exceptions illustrates how jurisdictions adapt to the realities of innovation. India must fashion its solutions to its context, protecting the vernacular content creators and small publishers as well as facilitate unencumbered innovation.

AI Authorship Dilemma

India’s Copyright Act still nests within anthropocentric presumptions. Section 2(d)(vi) grants authorship of computer-generated works to “the person who causes the work to be created,” which creates uncertainty over whether AI systems can qualify as authors. The General Clauses Act, 1897 defines “person” widely, but there is little direction in Indian jurisprudence regarding the extension of this status to algorithmic agents.

The key challenge lies in determining authorship and originality when machines independently generate creative works without direct human involvement. International precedents like Naruto v. Slater and the U.S. Copyright Office’s 2023 guidance reject copyright protection for non-human creations. India must clarify whether Section 17’s “first ownership” principle applies when the creator lacks legal personhood.

Conclusion

India urgently needs legislative reform to address the evolving AI–copyright interface. A comprehensive TDM exception framework, after the EU’s Articles 3 and 4 of the Digital Single Market Directive, could balance innovation and rights protection. Clear authorship guidelines should distinguish between human-assisted and autonomous AI creations, ensuring that legal personhood remains essential for ownership.

Additionally, standardized data licensing agreements and government-curated datasets can facilitate fair compensation and innovation. The Ministry of Commerce’s expert panel and the ANI Media v. OpenAI litigation together mark a transformative phase in India’s legal journey bridging the gap between analog era laws and algorithmic realities. India now stands at a pivotal moment to pioneer balanced governance that safeguards creativity while embracing technological progress.

[1] https://www.algindia.com/india-forms-expert-panel-to-review-ai-impact-on-copyright-law/

[2] CS(COMM) 1028/2024

[3] Section 52 of the Copyright Act, 1957