OWNING THE UNSEEN HAND: AUTHORSHIP, OWNERSHIP, AND THE LEGAL VACUUM SURROUNDING AI-GENERATED FASHION DESIGNS UNDER INDIAN AND INTERNATIONAL IP LAW

Authored By: RISHI VERMA

TEERTHANKER MAHAVEER UNIVERSITY, COLLEGE OF LAW AND LEGAL STUDIES, MORADABAD

ABSTRACT

Generative artificial intelligence has entered the fashion design studio as an autonomous aesthetic agent, producing garments and textiles that carry no conventional human touch. This article interrogates a critical vulnerability that Indian intellectual property jurisprudence is presently unequipped to resolve: when a commercial design is synthesised entirely by an algorithmic system, who owns it, and does any statutory protection vest at all? Through doctrinal analysis of Indian IP statutes and comparative engagement with the United States and European Union frameworks, this article demonstrates that the Indian IP framework is tethered to an authorship paradigm premised on human cognition, rendering existing law structurally incapable of accommodating fully autonomous creative outputs.

Keywords: Artificial Intelligence; Fashion Law; Luxury IP; Indian Copyright Law; Designs Act 2000; Algorithmic Authorship; AI Ownership; Comparative Intellectual Property

INTRODUCTION

Consider a garment that no human hand ever sketched. Its silhouette was generated by a diffusion model processing decades of runway archives; its textile pattern was composed by a neural network drawing from thousands of prior designs; its colourway was selected by a predictive analytics system calibrated to consumer data. The piece is commercially viable, aesthetically distinct, and production-ready. Yet when the question of ownership arises, Indian intellectual property law produces no answer because every statutory mechanism it possesses was built around one assumption: that somewhere behind the design there is a human creator.

This is no longer a hypothetical. Fashion enterprises across India and internationally are integrating generative AI systems into commercial design pipelines. India’s design houses, heritage textile sectors, and digital retail infrastructure are being reshaped by this technology at a pace that the legislative framework has not matched. The Copyright Act 1957, the Designs Act 2000[1], and the Trade Marks Act 1999[2] were each drafted on the premise of human creative agency. None contemplates a process in which aesthetic decisions are made autonomously by a machine.

This article argues that the resulting legal vacuum produces three failures. First, Indian law cannot identify a cognisable human author in a purely AI-generated design, rendering existing authorship rules ineffective. Second, even where authorship can be nominally attributed, no principled mechanism exists to determine ownership among the developer, deployer, and user of the AI system. Third, AI-generated outputs cannot reliably satisfy the novelty and originality requirements embedded in copyright and design law. Through comparative analysis with the United States and the European Union, this article proposes a tiered legislative framework that preserves the human-centred foundation of Indian IP law while creating workable rules for autonomous creative production.

BACKGROUND AND CONCEPTUAL FRAMEWORK

2.1 The Architecture of Fashion IP Protection in India

Fashion designs in India navigate three legislative regimes. The Copyright Act 1957 protects original artistic works, encompassing designs constituting works of artistic craftsmanship and original textile prints. The Designs Act 2000 provides registered protection for aesthetic features of industrially applied articles, subject to novelty and originality requirements. The Trade Marks Act 1999 protects distinctive brand identifiers, including non-conventional marks such as colour and product configuration, that distinguish one trader’s goods from another’s.

India’s obligations under the TRIPS Agreement[3] require protection for independently created industrial designs that are new or original. TRIPS is silent on authorship requirements, leaving that question entirely to domestic legislative choice. The Indian legislature therefore has both the freedom and the responsibility to address AI authorship through domestic law.

2.2 What Generative AI Does and Why It Matters Legally

Generative AI systems operate through a process fundamentally different from the tools Indian IP legislation contemplated. Trained on vast corpora of existing creative works, they generate outputs by sampling from learned probability distributions over aesthetic space. The resulting design is technically new but is not the product of deliberate human creative judgment.

What Indian IP law requires before protection vests is where this becomes legally consequential. Both copyright and design protection rest on the premise that a human mind originated the work and that a proprietor’s creative contribution grounds the claim to exclusive rights. Generative AI upends both premises. The Copyright Act 1957 defines authorship exclusively through human roles under section 2(d)[4]; the Designs Act 2000 vests rights in a proprietor construed as a natural or legal person under section 2(g)[5]. Neither provision accommodates a machine originator. This conclusion finds support in Thaler v Perlmutter No 1:22-cv-01564 (DDC 2023)[6], where the US District Court grounded its refusal to recognise AI authorship in the requirement that a work reflect human intellectual personality.

LEGAL ANALYSIS

3.1 The Copyright Framework: Where Section 2(d)(vi) Breaks Down

Section 2(d)[7](vi)[8] of the Copyright Act 1957[9] designates the author of a computer-generated work as the person who causes the work to be created. Introduced by the Copyright (Amendment) Act 1994[10] for human-operated software tools, this provision does not translate to generative AI. A user submitting a brief prompt exercises no meaningful directorial control; the system autonomously navigates parameter configurations and delivers output shaped by statistical patterns from its training data. To describe such a person as causing the work to be created stretches the provision beyond its rational interpretive limit.

The Supreme Court in Eastern Book Company v D B Modak (2008) 1 SCC 1[11] held that copyright requires skill and judgment reflecting independent intellectual effort such that the work carries the stamp of the author’s personality. Where creative selections are made by the AI system, the human contribution will rarely satisfy this threshold. Attribution under section 2(d)(vi) is therefore achievable, but protection is not. This produces a logically incoherent result.

Section 17[12] compounds this by vesting first ownership in the author without providing any mechanism to resolve competing claims among the AI developer, the contributors whose works trained the model, and the end user. This creates a tripartite ownership problem the statute was not designed to address.

3.2 The Designs Act 2000: Registration Without an Author

The Designs Act 2000 requires a proprietor in whose name a design may be registered. Because an AI system cannot hold legal rights, uncertainty arises as to whether ownership should vest in the developer, deployer, or user of the system. This uncertainty undermines the commercial reliability of any registration obtained for AI-generated designs.

The novelty requirement presents an additional challenge. An AI-generated design may appear original while incorporating elements derived from pre-existing works contained within training datasets. The Act provides no guidance on how novelty should be assessed in such circumstances. As of June 2026, the Office of the Controller General of Patents, Designs and Trade Marks[13] has issued no specific guidance on the registrability of AI-generated designs, leaving a significant regulatory gap.

3.3 Trade Marks as a Partial Workaround and Its Limits

Trademark law offers a limited solution when an AI-generated design acquires distinctiveness through association with a particular brand. Indian courts have recognised non-conventional marks[14], including colour and product configuration, demonstrating the flexibility of trademark protection in appropriate cases.

However, trademark law cannot substitute for copyright or design protection. It protects source identification rather than creative expression, arises only after distinctiveness has been established through use, and does not resolve ownership at the point of creation. While trademark protection may secure the commercial identity of an AI-generated design, it cannot address the underlying authorship and ownership questions created by generative AI.

CASE LAW DISCUSSION

4.1 Eastern Book Company v D B Modak (2008) 1 SCC 1[15]: India’s Originality Standard

In Eastern Book Company v D B Modak, the Supreme Court held that copyright protection requires skill and judgment reflecting independent intellectual effort rather than mere labour or investment. The Court rejected the “sweat of the brow” approach and adopted an originality standard grounded in human intellectual contribution. Applied to AI-generated fashion designs, the decision indicates that outputs lacking meaningful human creative involvement are unlikely to satisfy India’s originality requirement for copyright protection.

4.2 Cofemel[16]: Sociedade de Vestuário SA v G-Star Raw CV (C-683/17) [2019] — The EU on Fashion and Copyright

In Cofemel, the Court of Justice of the European Union confirmed that fashion designs may receive copyright protection provided they constitute original works reflecting the author’s own intellectual creation. The decision established that functional or applied artistic works are not excluded from protection merely because of their commercial nature. However, the requirement of intellectual creation remains tied to human authorship, making the judgment significant for assessing the protectability of AI-generated fashion designs under contemporary copyright standards.

4.3 Thaler v Perlmutter[17], No 1:22-cv-01564 (DDC 2023): The US Judicial Position

In Thaler v Perlmutter, the United States District Court upheld the Copyright Office’s refusal to register an AI-generated artwork, holding that human authorship remains a constitutional and statutory requirement of American copyright law. The decision confirms that purely autonomous AI outputs fall outside copyright protection. Read together with Naruto v Slater[18] and the Copyright Office’s 2023 guidance[19], the American position supports a distinction between fully AI-generated works and human-AI collaborative creations, with protection depending on the degree of human creative contribution.

CRITICAL ANALYSIS AND FINDINGS

5.1 The Structural Inadequacy of the Indian Framework

Indian IP law fails the AI authorship problem at three levels. Under copyright law, section 2(d)(vi) permits attribution without resolving whether protection should follow. Under design law, the Act cannot clearly identify a proprietor or assess novelty in relation to AI-generated outputs. Under trademark law, protection arises only after distinctiveness has been acquired through use, leaving ownership at the point of creation unresolved.

This uncertainty disproportionately affects smaller actors. Large fashion enterprises can manage risk through contractual allocation and brand development, whereas independent designers and creative studios often cannot. As noted by WIPO[20], unresolved AI authorship questions are likely to place smaller creative enterprises at a structural disadvantage.

5.2 Comparing Three Approaches and What India Should Learn

The United States and European Union exclude purely AI-generated works from copyright protection because both systems require human authorship. While the United States relies on constitutional and judicial authority, the European approach is grounded in the requirement of intellectual creation by a human author.

The United Kingdom adopts a more flexible approach through section 9(3)[21] of the Copyright, Designs and Patents Act 1988, which attributes authorship to the person making the arrangements necessary for creation. However, its application to modern generative AI remains uncertain. None of these jurisdictions has fully resolved the treatment of human-AI collaborative works. India therefore has the opportunity to develop a clearer framework that preserves human authorship while accommodating commercially significant AI-assisted design.

5.3 A Proposed Framework: Three Tiers

A tiered legislative framework provides the most coherent response to the authorship and ownership challenges created by generative AI.

For purely AI-generated designs produced without meaningful human creative contribution beyond a generic prompt, no intellectual property protection should vest and the design should enter the public domain upon publication. This approach preserves the public domain and avoids granting exclusive rights solely on the basis of technological ownership.

For human-AI collaborative designs involving substantial human creative control and aesthetic judgment, copyright and design protection should remain available. Section 2(d)(vi)[22] should be clarified to require a demonstrable degree of creative control, assessed through factors such as the specificity of instructions, the extent of human modification, and the nature of the aesthetic

decisions made during the creative process.

A third tier should address commercially deployed enterprise AI systems. Where no individual human creator can reasonably be identified, ownership should vest in the enterprise that deployed and directed the system through legislative assignment. Such protection should be subject to a fixed and shorter term than conventional copyright. This approach recognises commercial investment in AI-assisted design while preventing long-term monopolisation of machine-generated creative output.

CONCLUSION

The hand that operates a generative AI design system is, in a literal sense, unseen in the resulting design. Indian intellectual property law, as presently constituted, cannot trace that hand far enough to ground a coherent claim of authorship. The Copyright Act’s section 2(d)(vi), the Designs Act’s registration framework, and the Trade Marks Act’s distinctiveness doctrine each fail, at different points and for different reasons, to produce workable answers to the ownership questions that AI-generated fashion design raises.

The comparative analysis with the United States and the European Union confirms that this failure is not unique to India but that both jurisdictions have moved further toward legislative and administrative resolution. India must now follow and improve upon those approaches. The tiered framework proposed in this article, consisting of public domain treatment for purely AI-generated designs and calibrated protection for human-AI collaborative output, offers a principled and practically workable path forward. It preserves human authorship at the doctrinal centre of Indian IP law, aligns with India’s international obligations, and creates rules clear enough to be commercially useful for the designers, enterprises, and technologists who are building the AI-integrated future of Indian fashion. The unseen hand must, at some legislative moment, become visible. That moment is now.

REFERENCES AND BIBLIOGRAPHY

Table of Cases

Christian Louboutin SAS v Abubaker and Ors CS(COMM) 714/2016 (Delhi High Court, 2018)

Cofemel – Sociedade de Vestuário SA v G-Star Raw CV (C-683/17) [2019] (Court of Justice of the European Union)

Eastern Book Company v D B Modak (2008) 1 SCC 1 (Supreme Court of India)

Feist Publications Inc v Rural Telephone Service Co 499 US 340 (1991) (United States Supreme Court)

Hermès International v Mason Rothschild No 22-cv-384 (SDNY 2023)

Infopaq International A/S v Danske Dagblades Forening (C-5/08) [2009] ECR I-6569 (Court of Justice of the European Union)

Naruto v Slater 888 F 3d 418 (9th Cir 2018) (United States Court of Appeals for the Ninth Circuit)

Ritu Kumar v Biba Apparels (Delhi High Court)

Thaler v Perlmutter No 1:22-cv-01564 (DDC 2023) (United States District Court for the District of Columbia)

Thaler v Vidal 43 F 4th 1207 (Fed Cir 2022) (United States Court of Appeals for the Federal Circuit)

Table of Legislation

Designs Act 2000 (India)

Trade Marks Act 1999 (India)

Directive 2001/29/EC of the European Parliament and of the Council of 22 May 2001 on the harmonisation of certain aspects of copyright and related rights in the information society [2001] OJ L 167/10

Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 laying down harmonised rules on artificial intelligence [2024] OJ L 1689 (AI Act)

Table of Treaties

Agreement on Trade-Related Aspects of Intellectual Property Rights (Marrakesh, 15 April 1994)

Paris Convention for the Protection of Industrial Property (Paris, 20 March 1883)

Secondary Sources

Books and Legal Commentaries

Cornish W, Intellectual Property: Patents, Copyright, Trade Marks and Allied Rights (9th edn, Sweet and Maxwell 2019)

Journal Articles and Academic Papers

Scafidi S, ‘Intellectual Property and Fashion Design’ (2006) 1(1) Intellectual Property Law Review 115

Guadamuz A, ‘Artificial Intelligence and Copyright’ (2017) WIPO Magazine https://www.wipo.int/wipo_magazine/en/2017/05/article_0003.html accessed 5 June 2026

Official and Government Publications

European Parliament, ‘Report on Intellectual Property Rights for the Development of Artificial Intelligence Technologies’ (2020/2015(INI)) [2020]

NITI Aayog, National Strategy for Artificial Intelligence (Government of India 2018) https://www.niti.gov.in/national-strategy-artificial-intelligence accessed 5 June 2026

US Copyright Office, ‘Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence’ (March 2023) 88 Fed Reg 16190 https://www.federalregister.gov/documents/2023/03/16/2023-05321/copyright-registration-guidance-works-containing-material-generated-by-artificial-intelligence accessed 5 June 2026

WIPO, ‘Revised Issues Paper on Intellectual Property Policy and Artificial Intelligence’ (WIPO 2020) WIPO/IP/AI/2/GE/20/1 https://www.wipo.int/edocs/mdocs/mdocs/en/wipo_ip_ai_2_ge_20/wipo_ip_ai_2_ge_20_1.pdf accessed 5 June 2026

[1] Designs Act 2000.

[2] Trade Marks Act 1999 (India).

[3] Agreement on Trade-Related Aspects of Intellectual Property Rights (Marrakesh, 15 April 1994) art 25.

[4] Copyright Act 1957, s 2(d).

[5] Designs Act 2000, s 2(g).

[6] Thaler v Perlmutter No 1:22-cv-01564 (DDC 2023).

[7] Copyright Act 1957, s 2(d).

[8] Copyright Act 1957, s 2(d)(vi).

[9] Copyright Act 1957 (India).

[11] Eastern Book Company v D B Modak (2008) 1 SCC 1.

[12] Copyright Act 1957, s 17.

[13] Office of the Controller General of Patents, Designs and Trade Marks

[14] Christian Louboutin SAS v Abubaker and Ors CS(COMM) 714/2016 (Delhi High Court, 2018).

[15] Eastern Book Company v D B Modak (2008) 1 SCC 1 (Supreme Court of India).

[16] Cofemel – Sociedade de Vestuário SA v G-Star Raw CV (C-683/17) [2019] (Court of Justice of the European Union).

[17] Thaler v Perlmutter No 1:22-cv-01564 (DDC 2023).

[18] Naruto v Slater 888 F 3d 418 (9th Cir 2018).

[19] US Copyright Office, ‘Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence’ (March 2023) 88 Fed Reg 16190.

[20] WIPO, ‘Revised Issues Paper on Intellectual Property Policy and Artificial Intelligence’ (WIPO 2020) WIPO/IP/AI/2/GE/20/1.