In May 2025, the Ministry of Commerce & Industry, Government of India, constituted an eight‑member expert panel to evaluate whether the Copyright Act, 1957 (hereinafter, "the Act") adequately addresses the unique challenges presented by generative artificial intelligence (GenAI). The impetus was twofold: ongoing litigation—most prominently Ani Media (P) Ltd. v. Open AI Inc, 2024 SCC OnLine Del 8120, along with actions by the Federation of Indian Publishers, NDTV, Indian Express, and Hindustan Times alleging unauthorized use of copyrighted material in AI training, and the recognition that a statute enacted in the pre-digital era lacks explicit provisions for attributing authorship or ownership in the context of AI-generated works.
While the Ministry previously asserted that existing legal frameworks (including Section 52 of Copyright Act,1957 that states about the "fair dealing" exception) are sufficient to govern AI use , particularly insofar as commercial utilization of copyrighted works necessitates permission —the rapid proliferation of GenAI and related judicial scrutiny have propelled India toward concerted legal reform. The panel's mandate specifically includes:
- Analysing gaps in the Act relating to AI training, authorship, and ownership.
- Assessing whether storage, reproduction, and derivative outputs of copyrighted materials by GenAI constitute infringement.
- Recommending amendments or new provisions to ensure clarity and legal certainty.
This article proceeds to examine evolving jurisprudence, doctrinal principles, and comparative legal models, with a view to elucidating the current position and identifying potential reforms.
Understanding Generative AI and the Legal Challenge of Copyright Ownership
Generative AI, as defined under contemporary technological jurisprudence, refers to systems capable of producing text, images, code, and other outputs that resemble human creativity. These systems including models such as OpenAI's GPT-4, Google's Gemini, and Anthropic's Claude are trained on a massive corpora of data, much of which is copyrighted material scraped from the internet.
In India, copyright law is primarily governed by the Copyright Act, 1957, which recognizes copyright in original literary, artistic, musical, and dramatic works, and attributes ownership and authorship to a natural or juristic person. The term "author" is defined in Section 2(d) of the Act, and in the case of literary works, refers to the person who creates the work.
However, generative AI introduces a critical legal vacuum: Who is the "author" of a work autonomously generated by an AI system? Indian copyright jurisprudence, thus far, does not recognise non-human authorship. This creates three unresolved questions:
Can AI be considered an author?
Under current law, the answer is no. Indian copyright law, like its UK and US counterparts, limits authorship to persons—natural (humans) or juristic (companies, institutions).
Does the AI developer own the work?
If a software company trains a GenAI model that autonomously generates a poem or design, the company may assert ownership on the grounds of having created the tool. Yet, this claim lacks clear statutory backing in Indian law unless the output is traceably programmed or controlled.
What about user input?
When a human prompts an AI tool, does that prompt count as sufficient creative input to establish authorship? Courts have not yet adjudicated this in India, but under the doctrine of "minimal creativity," even small creative contributions may suffice for copyright protection, if attributable to a human.
Thus, Indian law presently does not recognize AI-generated content as copyrightable unless a human can be identified as its author or controller. This lacuna has led the expert panel to reconsider the scope of Sections 13 and 17 of the Copyright Act, 1957, which deal with what works are eligible for copyright and who holds first ownership.
Judicial Trends and Regulatory Signals in India
Although courts in India have not yet ruled definitively on generative AI and copyright ownership, recent judicial and regulatory developments signal growing concern and attention toward the issue.
A. Judicial Signals: ANI v. OpenAI (2024)
In a landmark 2024 suit before the Hon'ble Delhi High Court, Ani Media (P) Ltd. v. Open AI Inc, 2024 SCC OnLine Del 8120, ANI alleged that OpenAI's large language models (LLMs) used ANI's copyrighted news content for training purposes without a license or consent. ANI's petition invoked Sections 51 (infringement) and 55 (civil remedies) of the Copyright Act, contending that training AI with copyrighted material—even without direct reproduction in final output violated its exclusive rights.
The Delhi High Court admitted the matter and issued notice to the respondents, while emphasizing that copyright law does not currently provide a safe harbour or explicit training exemptions for GenAI developers. The case is ongoing but is widely viewed as a watershed moment in Indian copyright jurisprudence.
B. Regulatory Guidance: DPIIT's Position
The Department for Promotion of Industry and Internal Trade (DPIIT), under the Ministry of Commerce, clarified in mid-2024 that AI developers must seek authorization to use copyrighted content for training purposes, aligning with international best practices and reinforcing India's adherence to the Berne Convention.
Notably, the DPIIT rejected any blanket fair use defence under Section 52 of the Copyright Act, 1957 for commercial-scale AI training. Their position reflects growing alignment with the European Union Artificial Intelligence Act and United Kingdom Intellectual Property Office's consultation papers, which distinguish between private use and systemic data mining for profit-making purposes.
C. Expert Panel Recommendations (Preliminary Draft 2025)
As per leaked insights from the panel's early deliberations (published by The Print and Asia IP Law in June 2025):
- The panel is considering introducing a new Chapter (possibly Chapter XII-A) in the Act, specifically dealing with AI-generated works.
- Proposed changes include:
- Defining "AI-assisted" vs "AI-generated" works.
- Attributing authorship to the person who causes the work to be created, similar to Section 9(3) of the UK's Copyright Designs and Patents Act, 1988.
- Introducing mandatory disclosure norms for datasets used in training GenAI models.
- Establishing a statutory licensing scheme or levy for use of copyrighted works in training.
These developments suggest that copyright law is undergoing a foundational shift to accommodate emerging AI realities, while balancing creators' rights and innovation.
Comparative Legal Perspectives
India's ongoing efforts to reform copyright law vis-à-vis generative AI are part of a broader international discourse. By examining how other jurisdictions have approached the question of AI-generated works and copyright ownership, Indian lawmakers can draw informed, legally sound parallels.
A. United Kingdom
The UK's Copyright, Designs and Patents Act, 1988, is among the few statutes globally that explicitly address non-human authorship. Section 9(3) states:
"In the case of a literary, dramatic, musical or artistic work which is computer-generated, the author shall be taken to be the person by whom the arrangements necessary for the creation of the work are undertaken."
This provides a functional attribution model, wherein ownership rests with the human or corporate entity responsible for orchestrating the AI output. However, UK courts have yet to interpret this provision in the context of large-scale generative AI.
B. United States
Under U.S. law, copyright protection is firmly limited to human authorship. In Naruto v. Slater, the Ninth Circuit held that animals cannot claim copyright, setting a precedent that non-humans, including AI, have no authorship rights.
More recently, the U.S. Copyright Office (USCO) issued guidance (March 2023) stating that works generated "without any creative input or modification by a human" do not qualify for copyright protection. The USCO has rejected multiple copyright applications for purely AI-generated images or texts.
Nonetheless, the U.S. Congress is actively exploring a regulatory framework, with bills under consideration to establish licensing schemes for AI training data and possible attribution rights for creators whose work feeds into GenAI models.
C. European Union
The EU AI Act, passed in 2024, takes a multi-pronged approach. While not directly amending EU copyright laws, it:
- Mandates transparency in AI training datasets.
- Requires AI developers to respect copyright and related rights under the InfoSoc Directive.
- Encourages licensing models through Collective Management Organisations (CMOs) for text and data mining.
Additionally, the EU Parliament's draft AI copyright resolution proposes recognising the role of human input in AI-assisted works and creating a clear liability framework for infringement by GenAI outputs.
These comparative models offer a range of legislative and interpretive approaches—from outright denial of copyright to AI (U.S.) to conditional human attribution (UK and EU). For India, a hybrid attribution and liability model—balancing user input, developer responsibility, and rights of original creators—may be most appropriate.
Conclusion
India stands at a crucial inflection point in shaping how copyright law responds to the evolution of generative artificial intelligence. The Copyright Act, 1957, while progressive for its time, was drafted in an analogue era and does not adequately address the complexities of authorship, liability, and infringement in a digital, machine-learning-driven world.
The recently convened expert panel offers an important opportunity to modernise Indian copyright jurisprudence. Given the growing volume of litigation, India must act swiftly to provide predictable legal outcomes for creators, AI developers, and users alike. A balanced, innovation-friendly and rights-respecting approach will position India as a credible global leader in navigating the intersection of copyright and generative AI.
FAQ
1. Who owns copyright in AI-generated content under Indian law?
Currently, copyright law does not recognize AI as an author. Therefore, no copyright subsists in purely AI-generated content unless a human can be shown to have exercised creative control over the output. Ownership may vest in the user or developer only where significant human authorship is established.
2. Can training AI models on copyrighted data be considered copyright infringement in India?
Yes, if AI developers use protected works for training without authorization, it may amount to infringement under Section 51 of the Copyright Act, 1957, unless covered by a statutory exception or license. The DPIIT has clarified that such commercial use is not protected under fair dealing (Section 52).
3. Is there any legal guidance from Indian courts on generative AI and copyright?
Yes. The Delhi High Court's ongoing case—ANI v. OpenAI (2024)—is the first major judicial scrutiny of generative AI in India. While no final judgment has been delivered, the case may shape how courts view liability for AI training and authorship.
4. Will the Indian Copyright Act be amended to address AI-generated works?
Yes, the Government of India formed an expert panel in 2025 to review and modernise the 1957 Act. Proposals include defining AI-generated works, clarifying ownership, and introducing accountability mechanisms for AI use.
5. Can AI-generated content be protected by copyright if a human gives the prompt?
Possibly. If the human's input (prompt) demonstrates original creativity and control over the final output, the content may be considered an AI-assisted work, with copyright vesting in the human prompter. Courts will likely assess this on a case-by-case basis.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.