ARTIFICIAL INTELLIGENCE AND FASHION DESIGN: WHO OWNS AI-GENERATED FASHION?

Authored By: Fiona Anoop

Cochin University Of Science and Technology

ABSTRACT

The integration of artificial intelligence into the fashion design process has generated a profound legal vacuum concerning intellectual property ownership. This article examines who holds copyright and design rights over garments and textile patterns created by AI systems, with primary reference to Indian intellectual property law and the framework of the Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS). Through doctrinal analysis and examination of comparative jurisprudence from the United States, the United Kingdom, and the European Union, the article proposes targeted legislative reforms and interim judicial interpretations to bridge this gap, ensuring that the fashion industry’s rapid technological evolution is matched by proportionate legal certainty.

Keywords: Trade-Related Aspects of Intellectual Property Rights (TRIPS), IP Law, Artificial Intelligence, the Berne Convention

I. INTRODUCTION

In recent years, the traditional model of human-led design has been fundamentally disrupted by generative artificial intelligence. Tools such as Midjourney, DALL-E, Adobe Firefly, and purpose-built platforms like Stitch Fix’s algorithmic design engine now produce detailed garment illustrations, textile patterns, and complete seasonal collections within seconds. Major luxury houses including Gucci, Balenciaga, and Burberry have begun integrating AI-assisted workflows into their creative pipelines, while independent designers and fast-fashion aggregators alike employ AI to accelerate production cycles.

This transformation raises an urgent and largely unresolved legal question: when an AI system generates a fashion design, who owns the resulting intellectual property? Ownership of IP over a design governs who may exploit it commercially, who may prevent copying, and who receives royalty income. If AI-generated designs are owned by no one, they enter the public domain instantly, eliminating the incentive structures that IP law is designed to create. Conversely, if they are owned by AI developers, fashion brands using such tools may find themselves licensing rather than owning the products of their own creative processes.

II. BACKGROUND AND CONCEPTUAL FRAMEWORK

Intellectual property protection for fashion designs in India operates primarily under two legislative instruments: the Copyright Act, 1957 and the Designs Act, 2000. The Copyright Act, 1957 protects “artistic works” under Section 13, which includes paintings, drawings, engravings, and works of artistic craftsmanship. Textile prints, surface patterns, and two-dimensional design elements may qualify as artistic works if they meet the threshold of originality, following the Supreme Court’s landmark ruling in Eastern Book Company v. D.B. Modak.

The Designs Act, 2000 separately protects the “features of shape, configuration, pattern, ornament or composition of lines or colours” applied to an article through an industrial process, provided the design is novel and not previously published. Importantly, once an artistic work is applied industrially more than fifty times, it loses copyright protection and may only be protected under the Designs Act, as per Section 15 of the Copyright Act, 1957.

At the international level, the TRIPS Agreement, to which India is a signatory, establishes minimum standards of IP protection. Article 9 of TRIPS incorporates Articles 1 through 21 of the Berne Convention, requiring member states to protect literary and artistic works, including applied arts. Article 2(1) defines protectable works as “productions in the literary, scientific and artistic domain,” language historically understood to contemplate human creativity. Prior scholarship by scholars such as Ryan Abbott and Andres Guadamuz has proposed frameworks ranging from sui generis protection to compulsory licensing, but no consensus has emerged globally.

III. LEGAL ANALYSIS

A. The Human Authorship Requirement in Indian Copyright Law

In the Copyright Act, 1957, section 2(d) defines an author in relation to a literary or artistic work as the person who creates it. The word “person” in Indian legal usage ordinarily refers to a legal person. A company is a legal person and may thus be the owner, if not the creator, of copyright. However, the Copyright Act, 1957 distinguishes between authorship and ownership. Section 17 vests first ownership of copyright in the author, subject to exceptions for works made in the course of employment.

Given that AI lacks legal personhood in Indian law, it cannot be an author. This was implicitly affirmed in the United States when the Copyright Office refused registration for the AI-generated artwork “A Recent Entrance to Paradise” in Thaler v. Vidal, holding that copyright requires human authorship. The Delhi High Court in Anil Gupta v. Kunal Dasgupta emphasised that copyright subsists in works reflecting the “intellectual creation” of their maker.

Section 17 of the Copyright Act, 1957 addresses works made “by an author in the course of employment” but says nothing about works generated by a tool used by an employee. The closest analogy is the “computer-generated work” provision found in Section 9(3) of the UK Copyright, Designs and Patents Act, 1988, which vests authorship in “the person by whom the arrangements necessary for the creation of the work are undertaken.” India has no equivalent provision.

B. Design Protection under the Designs Act, 2000

Section 2(g) defines “proprietor of a new or original design” as the person who creates it or acquires it from the creator. The Act is applicant-centric for registration purposes. This creates an opening: a fashion company that deploys an AI system to generate a novel design pattern, registers that design, and applies it industrially could theoretically secure design protection. The “novelty” and “originality” requirements under Section 4 of the Designs Act, 2000 do not expressly require human creative input.

However, this opening is narrow. Design registration under the Designs Act, 2000 requires that the design be novel or original, not previously published anywhere in the world. Given that AI systems trained on vast fashion datasets may generate designs with considerable statistical similarity to existing registered designs, the risk of inadvertent novelty failure is substantial.

C. The TRIPS Framework and Its Implications

TRIPS Article 27 requires member states to make patents available for “any inventions, whether products or processes, in all fields of technology,” provided they are new, involve an inventive step, and are industrially applicable. The question of whether AI can be an inventor for patent purposes, addressed in Thaler v. Comptroller-General of Patents by the UK Supreme Court, held that an AI system cannot be named as an inventor under the Patents Act, 1977 because the Act requires inventors to be “natural persons.”. India could therefore legislate to accommodate AI-generated works without violating TRIPS, provided minimum protection standards are maintained.

TRIPS Article 10(2) extends protection to compilations of data as intellectual creations “by reason of the selection or arrangement of their contents.” Some scholars argue that the output of an AI trained on fashion design data constitutes a form of compilation. This argument is strained, because compilations protect the selection and arrangement, not the underlying materials and the fashion design itself, does not neatly fit the compilation paradigm.

A further complication arises from TRIPS Article 39 on the protection of undisclosed information. Fashion houses increasingly treat their AI training datasets as trade secrets. This avenue of protection does not resolve the ownership of AI outputs, but it provides a parallel layer of exclusivity over the inputs, creating indirect competitive advantage even in the absence of output protection.

D. Moral Rights and AI-Generated Fashion

Section 57 of the Copyright Act, 1957 confers upon the author, independently of copyright, the right of paternity and the right of integrity, including the right to object to distortion or mutilation of the work prejudicial to honour or reputation. These rights persist even after the copyright is transferred. If AI-generated fashion designs are treated as having no human author, Section 57 is inapplicable, eliminating this protection entirely. The consequence is that an AI-generated garment design can be modified, distorted, or attributed falsely without any legal remedy which is an outcome that undermines the credibility of AI-assisted fashion brands.

IV. CASE LAW DISCUSSION

A. Thaler v. Vidal, 43 F.4th 1207 (Fed. Cir. 2022)

Dr. Stephen Thaler sought to register patents in the United States for inventions he attributed to an AI system called DABUS (Device for the Autonomous Bootstrapping of Unified Sentience). The United States Court of Appeals for the Federal Circuit held that the Patent Act requires inventors to be natural persons, and that DABUS could not be an inventor. The court reasoned that the Act consistently uses personal pronouns such as “himself” and “herself”, to describe inventors, reflecting congressional intent to confine inventorship to human beings. The same reasoning has been applied in analogous copyright proceedings before the US Copyright Office, which has repeatedly refused to register AI-generated works lacking human authorship. The legal principle established is clear: under laws designed for human creators, AI systems cannot be rights-holders. Its relevance to the present inquiry is direct as it confirms the global trend of restricting creative rights to humans, and underscores the lacuna in Indian law, which lacks even the procedural mechanisms to definitively refuse AI authorship.

B. Thaler v. Comptroller-General of Patents [2023] UKSC 49

Before the UK Supreme Court, Dr. Thaler again sought patent recognition for DABUS-generated inventions. The Supreme Court unanimously held that the UK Patents Act, 1977 requires an inventor to be a person, and that a machine cannot hold the right to be granted a patent. Crucially, the Court observed that this conclusion was compelled by statutory interpretation and left open the normative question of whether the law ought to be reformed to accommodate AI inventorship. The case is directly relevant to Indian fashion law because India’s IP statutes similarly contain no AI-specific provisions. It demonstrates that without legislative intervention, courts are likely to hold that AI cannot be an author or proprietor under existing Indian law, regardless of how commercially significant AI-generated designs may be.

C. Telstra Corporation Limited v. Phone Directories Company Pty Ltd [2010] FCAFC 149

In this Australian Federal Court decision, the Full Court held that copyright does not subsist in telephone directories compiled through automated processes without sufficient human creative expression. The Court applied the principle that copyright requires intellectual effort from a human author. Where the creative choices are made by a computer program without human selectivity or judgement at the level of individual entries, copyright does not arise. While Australian law is a common law jurisdiction distinct from India, the Telstra decision has persuasive value in any common law copyright analysis. Applied to AI-generated fashion design, it supports the conclusion that a garment or textile pattern produced entirely by an AI system, would not attract copyright protection in India. This creates a significant commercial risk.

V. CRITICAL ANALYSIS AND FINDINGS

The foregoing analysis reveals three systemic gaps in Indian IP law as applied to AI-generated fashion design. First, there is an authorship vacuum: since AI cannot be an author under the Copyright Act, 1957, and Indian law contains no computer-generated works provision equivalent to Section 9(3) of the UK CDPA, AI-generated fashion designs fall outside the copyright framework entirely.

Second, there is a design registration gap. While the Designs Act, 2000 does not explicitly require human authorship for registration, the scale at which AI generates designs overwhelms traditional registration procedures. The requirement of individual examination for novelty becomes unworkable when a single AI session produces thousands of candidate designs. India’s IP Office lacks the infrastructure and procedural rules to handle AI-generated design applications systematically.

Third, there is a moral rights paradox. India’s strong moral rights tradition under Section 57 of the Copyright Act, 1957 is premised on human dignity and creative integrity. Its inapplicability to AI-generated works means that a human designer who trains an AI on their own creative history may lose moral rights over derivative outputs which is a perverse outcome that could discourage human designers from contributing to AI training datasets.

Comparatively, the European Union has taken a more proactive approach. The EU AI Act, 2024 establishes transparency obligations for AI-generated creative content, while the European Commission has signalled intent to address IP ownership of AI outputs through the Intellectual Property Action Plan. India, by contrast, lacks any AI-specific IP policy framework. The Office of the Controller General of Patents, Designs & Trade Marks has not issued guidelines on AI-generated works, and no Indian appellate court has directly addressed the question.

VI. CONCLUSION

This article has demonstrated that Indian intellectual property law, as currently constituted, does not provide adequate protection or ownership rules for AI-generated fashion designs. The Copyright Act, 1957 requires human authorship, which AI cannot satisfy. The Designs Act, 2000 offers a partial avenue through its applicant-centric registration system, but administrative infrastructure is inadequate to the scale of AI design generation. India’s TRIPS obligations permit legislative reform but do not compel it.

Three reforms are recommended. First, the Copyright Act, 1957 should be amended to introduce a “computer-generated works” provision, vesting copyright in the person who makes the necessary arrangements for the AI to create the work. Second, the Designs Act, 2000 should be amended to introduce a streamlined bulk-registration procedure for AI-generated designs, with novelty assessed through automated comparison against an expanding prior art database. Third, the Copyright Act, 1957 should clarify that moral rights under Section 57 attach to the human individuals who contributed creative input to an AI’s training or prompt design, ensuring that India’s human-centric moral rights tradition is preserved even as the creative process evolves.

As AI reshapes the boundaries of fashion creativity, the law must respond with equal agility. The reforms proposed above would bring India to the frontier of global fashion law jurisprudence and safeguard both human creativity and technological innovation.

REFERENCES AND BIBLIOGRAPHY

Primary Sources

Legislation

Designs Act 2000 (India)

Trade Marks Act 1999 (India)

Patents Act 1970 (India)

Patents Act 1977 (United Kingdom)

EU Artificial Intelligence Act, Regulation (EU) 2024/1689 of the European Parliament and of the Council

International Instruments

Agreement on Trade-Related Aspects of Intellectual Property Rights, Annex 1C to the Marrakesh Agreement Establishing the World Trade Organization (15 April 1994) 1869 UNTS 299 (TRIPS Agreement)

Berne Convention for the Protection of Literary and Artistic Works (9 September 1886, as revised) 1161 UNTS 30 (Berne Convention)

Paris Convention for the Protection of Industrial Property (20 March 1883, as revised) 828 UNTS 305

Cases

Anil Gupta v. Kunal Dasgupta (2002) 25 PTC 1 (Delhi HC)

Eastern Book Company v. D.B. Modak (2008) 1 SCC 1 (SC India)

Thaler v. Comptroller-General of Patents, Designs and Trade Marks [2023] UKSC 49

Thaler v. Vidal, 43 F.4th 1207 (Fed Cir 2022)

Telstra Corporation Limited v. Phone Directories Company Pty Ltd [2010] FCAFC 149 (Full Fed Ct Australia)

Secondary Sources

Books

Abbott R, The Reasonable Robot: Artificial Intelligence and the Law (Cambridge University Press 2020)

Scafidi S, Who Owns Culture? Appropriation and Authenticity in American Law (Rutgers University Press 2005)

Waelde C and others, Contemporary Intellectual Property: Law and Policy (5th edn, Oxford University Press 2022)

Journal Articles

Ginsburg JC, ‘The Concept of Authorship in Comparative Copyright Law’ (2003) 52 DePaul L Rev 1063

Scafidi S, ‘Intellectual Property and Fashion Design’ (2007) 1 Intellectual Property and Information Wealth 115

McCutcheon J, ‘Intuitel and Copyright: Is There a Place for Artificial Intelligence in Intellectual Property Law?’ (2013) 36 UNSW Law Journal 1

Online Sources

World Intellectual Property Organization, ‘WIPO Conversation on Intellectual Property (IP) and Artificial Intelligence (AI)’ (WIPO 2020) <https://www.wipo.int/about-ip/en/artificial_intelligence/> accessed 7 June 2026

US Copyright Office, ‘Copyright and Artificial Intelligence, Part 2: Copyrightability’ (February 2025) <https://www.copyright.gov/ai/> accessed 8 June 2026

European Commission, ‘Intellectual Property Action Plan’ (November 2020) <https://ec.europa.eu/commission/presscorner/detail/en/ip_20_2187> accessed 8 June 2026