Published On: February 17th 2026
Authored By: Ajit Raju Kamble
Symbiosis Law School, Pune
Abstract
AI companies in India face significant legal uncertainty regarding the use of copyrighted material to train machine learning models. The absence of clear rules governing text-and-data mining (TDM), coupled with ongoing debates over creator compensation, creates substantial business and litigation risks that could impede innovation. This article examines how Indian copyright law applies to AI training and outputs, analyzes relevant judicial precedents, and compares India’s emerging framework with approaches adopted in the United States, European Union, and China.
I. Introduction
India is rapidly emerging as a global center for artificial intelligence development, yet its legal framework has not kept pace with technological advancement. Generative AI is fundamentally reshaping creative industries—transforming how we write, compose music, generate images, and produce code—but Indian copyright law remains rooted in principles designed for human creators. The Copyright Act, 1957, defines authorship and protection around human creativity, leaving wholly machine-generated works in legal limbo.[1]
The challenges facing AI companies in India are multifaceted. Large-scale model training frequently depends on vast collections of text, images, and other potentially protected works scraped from the internet. Meanwhile, Indian policymakers are grappling with whether new regulations—such as a national text-and-data-mining exception, collective licensing mechanisms, or mandatory blanket licenses—are necessary to balance innovation against creator rights.[2]
This article examines three core questions: First, does gathering, copying, or processing copyrighted works to train machine learning models constitute copyright infringement or permissible research activity? Second, who owns the copyright in AI-generated content? Third, how should India structure its legal framework to encourage innovation while protecting creator interests?
II. Relevant Statutory Definitions
Understanding AI copyright issues requires familiarity with key statutory concepts under the Copyright Act, 1957.
A. Authorship
Section 2(d) of the Copyright Act defines “author” differently depending on the type of work. For literary or dramatic works, the author is “the person who creates the work.” For photographs, it is “the person who takes the photograph.” For cinematograph films, it is “the producer.” Notably, for computer-generated works, the author is defined as “the person who causes the work to be created.”[3] This last provision creates interpretive space for determining authorship in AI contexts, though it assumes human causation.
B. Ownership
Section 17 prescribes that, subject to any agreement to the contrary, the author of a work is the first owner of copyright. Special ownership rules apply for works made by employees during employment, works created under commission, and works produced by government or international organizations.[4]
III. Understanding Artificial Intelligence
Artificial Intelligence (AI) refers to the capability of digital computers or computer-controlled systems to perform tasks typically associated with intelligent beings. The term encompasses systems endowed with intellectual processes characteristic of humans, such as reasoning, discovering meaning, generalizing, or learning from past experience.[5]
Since the 1940s, digital computers have been programmed to execute complex tasks—from proving mathematical theorems to playing chess at expert levels. Despite advances in processing speed and memory capacity, no programs yet match full human flexibility across diverse domains or tasks requiring extensive everyday knowledge. However, some programs have achieved performance levels comparable to human experts in specific tasks, making AI ubiquitous in applications ranging from medical diagnosis and search engines to voice recognition and chatbots.[6]
IV. Legal Challenges Facing AI Companies in India
AI companies operating in India confront numerous legal obstacles stemming from the absence of AI-specific legislation and the uncertain application of existing laws to novel AI scenarios.
A. Absence of AI-Specific Legislation
India currently lacks unified, dedicated legislation addressing AI’s unique characteristics. Existing laws cover general data protection, cybersecurity, and intermediary liability, but none are specifically tailored to algorithmic decision-making, opaque models (the “black box” problem), autonomous systems, self-learning algorithms, or deepfakes. Without clear legal standards for AI liability, transparency obligations, and fairness norms, companies face regulatory uncertainty.
B. Transparency, Explainability, and Algorithmic Bias
AI systems can perpetuate or amplify biases present in training data, including biases related to gender, caste, race, and socioeconomic status. When algorithms operate without transparency, affected individuals find it difficult to contest decisions or seek redress. The lack of explainability requirements compounds these problems.
C. Privacy and Data Protection
AI systems typically require large datasets that may contain sensitive personal information and biometric data. Despite the enactment of the Digital Personal Data Protection Act, 2023, significant issues remain unresolved, including standards for data minimization, purpose limitation, anonymization versus pseudonymization, cross-border data transfers, exceptions for state and law enforcement authorities, and the potential for mass surveillance or misuse.[7]
D. Liability and Accountability
When AI systems cause harm—through wrongful decisions, discrimination, or physical injury in autonomous systems—questions arise about who bears responsibility. Is it the developer, the deployer, the user, or the AI system itself? The absence of clear liability frameworks creates accountability gaps that can deter responsible innovation or allow harmful actors to evade consequences.
E. Intellectual Property Issues
AI-generated content—including art, literature, code, and music—raises fundamental copyright questions. Current intellectual property laws in India do not satisfactorily address who owns copyright in works created autonomously by AI, or whether such works qualify for copyright protection at all. Additionally, the use of copyrighted training data implicates reproduction and adaptation rights, potentially exposing AI companies to infringement claims.
F. Ethical, Social, and Socioeconomic Implications
AI’s broader implications extend beyond legal compliance. Issues include labor displacement, exacerbation of inequality, threats to democratic discourse through deepfakes and disinformation, surveillance, and discrimination. There is also potential for chilling effects if users fear constant monitoring. Effective regulation must address not only legal liability but also ethical standards and social responsibility.
G. Enforcement, Technical Capacity, and Fragmentation
Regulations are only as effective as their enforcement mechanisms. Indian regulators may lack the technical expertise necessary to audit AI systems, test for bias, or validate algorithms. Furthermore, patchwork regulation—with different sectors subject to disparate rules in health, finance, and telecommunications—can produce inconsistent standards and compliance burdens.
H. Cross-Border and International Law Issues
AI systems and data frequently traverse jurisdictions. Training data may be stored overseas, models may be trained on international datasets, and outputs may be shared worldwide. This creates complex questions about which country’s laws apply and how to enforce judgments across borders.
V. Judicial Precedents on Copyright and Computer-Generated Works
Indian courts have addressed related copyright issues in cases predating modern AI, offering interpretive guidance for current challenges.
A. R.G. Anand v. Delux Films (1978)
The Supreme Court established that copyright protects the expression of ideas rather than ideas themselves.[8] This foundational principle creates interpretive space for determining what constitutes sufficient human expression in AI-assisted work. If a human user provides creative direction through prompting or iterative refinement, thereby substantially shaping the final work’s expression, copyright protection may vest in that human contribution even if AI executes the mechanical aspects of creation. However, purely autonomous AI outputs lacking human direction regarding expression likely remain unprotected under this framework.
B. Eastern Book Company v. D.B. Modak (2007)
The Supreme Court held that originality in copyright derives from skill and judgment requiring intellectual effort and contribution, rather than mere mechanical application of tools.[9] The Court determined that using computer technology to manipulate existing material without independent creative contribution fails to meet the originality threshold for copyright eligibility. Though this precedent predates contemporary AI, it establishes that algorithmic processing alone cannot constitute authorship.
This holding creates tension within the statutory framework: Section 2(d) extends authorship to “the person who causes” creation, yet offers no recognition to artificial entities operating substantially autonomously from human direction. The question remains whether prompting or training an AI model constitutes sufficient “causing” to establish authorship.
VI. Comparative Analysis: International Approaches to AI and Copyright
Examining how other major jurisdictions address AI copyright issues illuminates potential pathways for India.
A. United States
The United States maintains a traditional approach requiring human authorship for copyright protection while explicitly rejecting copyright for purely AI-generated materials. The U.S. Copyright Office has consistently stated that copyright law requires human creativity and that works produced entirely by machines or automated processes without human intervention are not copyrightable.[10] This position reflects the broader American preference for resolving emerging technology challenges through judicial interpretation of existing frameworks rather than new legislation.
Judicial precedents reinforce this stance. In Thaler v. Perlmutter, the U.S. District Court for the District of Columbia held that copyright requires human authorship, rejecting an attempt to register an AI-generated artwork.[11] The Copyright Office’s approach emphasizes that human contribution—through selection, arrangement, or modification of AI outputs—can render derivative works copyrightable even when the underlying AI-generated elements are not.
B. European Union
The European Union has developed a more interventionist regulatory framework emphasizing a “significant human input” standard for AI-generated content. Under this approach, AI-generated works may qualify for copyright protection only when human involvement demonstrates substantial contribution to the creative process.[12] The EU framework mandates comprehensive content labeling requirements and establishes clear liability frameworks for AI technology companies.
The EU’s AI Act, adopted in 2024, imposes transparency obligations, risk assessment requirements, and human oversight mandates for high-risk AI systems. This comprehensive regulatory approach reflects European preferences for ex-ante regulation to protect fundamental rights and establish clear legal boundaries for emerging technologies.
C. China
China presents the most progressive approach among major economies regarding AI copyright recognition. The Beijing Internet Court’s landmark decision in a November 2023 case established precedent by recognizing copyright protection for AI-generated images where demonstrable human intellectual effort was involved in the creative process.[13] The court emphasized that the plaintiff’s selection of prompts, parameters, and iterative refinements constituted sufficient creative input to warrant protection.
China’s framework requires AI-generated content to be clearly labeled and holds AI companies responsible for their systems’ actions. This approach creates balanced rules supporting technological development while protecting creator rights through effective coordination between government departments and industry stakeholders.
VII. Conclusion
India faces a critical juncture in developing its AI copyright framework. The rapid advancement of generative AI technology and the growth of India’s creative and technology industries demand immediate policy attention. While stakeholder consensus opposes standalone AI legislation that could constrain innovation, targeted regulatory interventions remain necessary to protect creator interests and provide legal certainty.
The current copyright regime requires updates to address AI-generated content complexity and the use of copyrighted works in training datasets. Key policy questions include: Should India adopt a text-and-data-mining exception similar to those in the EU and Japan? How should the law allocate rights between AI developers, users, and creators of training data? What transparency and labeling requirements should apply to AI-generated content?
A balanced regulatory approach is essential—one that protects creator interests while enabling the innovation crucial for India’s technological leadership. Industry organizations like NASSCOM advocate for broad TDM exemptions to safeguard innovation, while rights holders push for licensing or royalty systems to ensure fair compensation. Resolving this tension through thoughtful legislation and judicial interpretation will determine whether India can establish itself as a global AI hub while maintaining robust protections for creative industries.
Policymakers should consider: (1) clarifying authorship standards for AI-assisted works; (2) establishing fair use or TDM exceptions for training purposes; (3) creating transparency requirements for AI-generated content; (4) developing liability frameworks that encourage responsible innovation; and (5) fostering dialogue among technologists, creators, and legal experts to craft nuanced solutions. Only through such comprehensive reform can India navigate the complex intersection of AI advancement and copyright protection.
References
[1] The Copyright Act, 1957, No. 14 of 1957, INDIA CODE (1957).
[2] NASSCOM, Policy Recommendations on Artificial Intelligence and Copyright (2024).
[3] The Copyright Act, 1957, § 2(d).
[4] Id. § 17.
[5] Stuart Russell & Peter Norvig, Artificial Intelligence: A Modern Approach (4th ed. 2020).
[6] See generally Artificial Intelligence, ENCYCLOPAEDIA BRITANNICA, https://www.britannica.com/technology/artificial-intelligence (last visited Feb. 17, 2026).
[7] The Digital Personal Data Protection Act, 2023, No. 22 of 2023, INDIA CODE (2023).
[8] R.G. Anand v. M/S Delux Films & Ors., (1978) 4 SCC 118 (India).
[9] Eastern Book Company & Ors. v. D.B. Modak & Anr., (2008) 1 SCC 1 (India).
[10] U.S. Copyright Office, Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence, 88 Fed. Reg. 16,190 (Mar. 16, 2023).
[11] Thaler v. Perlmutter, 638 F. Supp. 3d 26 (D.D.C. 2023).
[12] European Parliament and Council Regulation 2024/1689 on Artificial Intelligence (AI Act), 2024 O.J. (L) (EU).
[13] Beijing Internet Court, Case No. (2023) Jing 0491 Min Chu 11279 (Nov. 27, 2023) (China).
