Authored By: Mrudula Kale
University of Mumbai
Abstract
The mix of artificial intelligence and fashion design raises a significant legal question in today’s digital world: when an algorithm designs a piece of clothing, who owns it? This article explores the legal issues caused by AI-generated fashion, focusing mainly on Indian intellectual property law and making comparisons to international laws. The main question is whether current copyright, design protection, and trademark laws can effectively include AI as a creative force in the fashion industry. By reviewing the Copyright Act 1957, the Designs Act 2000, and the Trade Marks Act 1999, along with key international cases and evolving global laws, this article concludes that Indian IP law is currently unable to recognise machine authorship. It calls for urgent legal reforms to fill the ownership gap created by AI-generated fashion, while also evaluating suggestions for a new framework tailored for AI authorship based on best practices from other countries.
I. Introduction
Imagine this: an AI system, trained on decades of Dior archives, Sabyasachi collections, and runway videos from Paris to Mumbai, creates a stunning silk evening gown all by itself. The gown features a unique hand-embroidered pattern that draws from no existing tradition, a silhouette that no human has ever envisioned, and a colour scheme designed entirely in binary. The gown is made, sold, and praised. Then, a lawsuit arises. But who is filing the claim — and against whom?
This scenario is no longer theoretical. AI systems like Google’s DeepDream, Stitch Fix’s design engine, and various generative adversarial networks (GANs) are producing visually appealing designs with real commercial use in fashion.[1] However, the intellectual property laws that oversee these creations were developed during a time when creativity was clearly human. Now those laws are showing signs of strain.
This article raises a straightforward question: can a machine qualify as a fashion designer under the law? Finding the answer involves navigating complex ideas about authorship, originality, and legal status across three areas of IP law. The article is organised as follows. Section II lays out the background and legal context. Section III analyses Indian laws and international frameworks in detail. Section IV reviews significant court decisions. Section V summarises key findings. Finally, Section VI presents conclusions and suggests reforms.
II. Background and Conceptual Framework
A. What Is AI-Generated Fashion?
AI-generated fashion includes garments, textile patterns, accessories, or complete design concepts created mainly by machine learning algorithms, neural networks, or other autonomous AI systems without direct human creative input during the generation process.[2] The range of creativity is vast: on one side, AI acts as a refined tool — like a digital sewing needle — completely controlled by human direction. On the opposite side, AI operates as an independent creative entity, producing results that even surprise its developers. The legally relevant category is the latter: outputs where the AI makes creative decisions about colour, shape, pattern, proportion, and decoration instead of the human operator. This area puts the strain on current IP laws, and it is the focus of this article.
B. The Doctrine of Authorship: A Distinctly Human Affair
Intellectual property law in major jurisdictions has consistently linked authorship to human identity. This connection is based on Lockean labour theory, which holds that individuals gain rights over what they create through their intellectual effort.[3] In Indian law, this principle is outlined in the Copyright Act 1957, which grants copyright protection to original artistic works.[4] This standard assumes that a human creator exercised their creative judgment. The Designs Act 2000 likewise focuses on human creativity, protecting designs that show novelty and originality, understood traditionally as outcomes of human aesthetic choices.[5] Neither this Act nor the Trade Marks Act 1999 considers the idea that a non-human entity could make creative contributions resulting in protectable works.[6]
C. AI and Legal Personality: The Conceptual Gap
Legal personality — the ability to hold rights and responsibilities — in Indian law applies to natural persons and recognised legal entities like companies, trusts, and even certain religious idols.[7] AI systems do not fit into either category; they cannot enter into contracts, bear responsibility, or own property. This gap means that when AI generates a valuable fashion design, the law does not have a clear solution to the ownership question, leading to what scholars have referred to as a “creativity vacuum.”[8]
D. Fashion’s Particular Vulnerability to the AI Authorship Problem
Fashion, more than other creative industries, faces distinct risks from AI authorship. The industry works with fast commercial cycles, heavily relies on protection for patterns and motifs, and derives significant value from the link between creative origins and brand identity.[9] AI disrupts the chain of human creativity, threatening the foundations of copyright protection, design registration, and brand authenticity — three cornerstones of fashion law. In brief, machines do not just change how fashion is created; they challenge the very legal basis for fashion’s commercial value.
III. Legal Analysis
A. Copyright Law and the Authorship Requirement
The Copyright Act 1957 is the main law that governs artistic works in India. Section 13 of the Act gives copyright protection to original artistic works.[10] The Supreme Court set the standard for originality in Indian law in Eastern Book Company v D.B. Modak,[11] adopting a standard of “minimal degree of creativity”, rejecting simple skill and effort as enough, and requiring that the author make real creative choices. This is a low threshold, but it is clearly a human one.
The Act defines “author” in Section 2(d) as the person whose intellectual creation a work represents.[12] Section 17, which deals with first ownership of copyright, gives ownership to the author or, for works created during employment, to the employer.[13] Neither of these provisions accounts for a situation where the creator is an autonomous algorithm with no employer, no employment relationship, and no legal identity.
The Act’s nearest attempt at handling computer-generated work is Section 2(d)(vi), which states that for computer-generated works, the author is “the person who causes the work to be created.”[14] This provision was written when word processors and basic graphics software were common. It aimed to address human-operated tools, not generative AI systems that create outputs without specific human guidance. In the context of AI fashion design, the “person who causes the work to be created” might be the AI developer, the end user who inputs a prompt, or the corporate operator who uses the system. This creates more confusion than clarity and rewards the loudest claim of authorship rather than true creators.
B. Design Protection Under the Designs Act 2000
The Designs Act 2000 protects the visual features of industrially applied articles.[15] For a design to be registered, it must be new or original[16] and must not have been shown to the public before registration.[17] Unlike copyright, design protection depends on registration; the applicant, not the creator, is the key legal concept.
This registration-based system offers a partial workaround for AI-generated designs. A human or corporate entity can, in theory, register an AI-generated design and gain exclusive rights simply by filing the necessary paperwork. However, this solution hides a legal difficulty. The purpose of the Designs Act is to reward creative investment. Allowing someone who merely uses an AI system to get the same protection as a human designer who spent months creating a garment misaligns the law’s intent with its application. The filing clerk replaces the creator. This situation is not acceptable — legally or morally.
C. Trade Mark Law and the AI Identity Problem
The Trade Marks Act 1999 protects signs that distinguish the goods or services of one entity from those of another.[18] In fashion, trade marks protect logos, brand names, and distinctive labels, as well as specific design elements, such as Louboutin’s famous red sole.[19]
The issue of AI authorship overlaps with trade mark law in an important way. Trade marks serve a legal purpose based on association; the mark tells the consumer something about the origin, quality, and identity of the product. When AI creates a design, the “origin story” — who made it and with what vision — is often missing or untraceable. This undermines the basic consumer-protection goal of trade mark law. Section 11(1) of the Trade Marks Act 1999 states that confusion about origin is a valid reason for refusing registration.[20] If an AI-generated design is marketed as a luxury brand’s signature piece, but no human creativity is behind it, the mark may itself constitute a misrepresentation. The law does not have a clear answer yet.
D. International Framework: TRIPS and the Berne Convention
India’s intellectual property obligations are influenced by the Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS).[21] TRIPS requires member states to provide copyright protection for literary and artistic works but does not define “author,” leaving that determination to domestic law.[22] The Berne Convention for the Protection of Literary and Artistic Works also requires protection for “authors” without defining authorship. This gap is understandable for a nineteenth-century agreement but leaves member states without guidance on the issue of AI.[23]
WIPO’s ongoing discussions about AI and intellectual property, including the 2019 WIPO Conversation on Intellectual Property and Artificial Intelligence, have produced extensive commentary but no binding international consensus.[24] This gap at the international level has forced local governments to come up with their own solutions, leading to inconsistent results. The technology works globally, but the law remains stuck in individual countries.
IV. Case Law Discussion
A. Eastern Book Company v D.B. Modak: The Indian Originality Baseline
Though not a fashion case, the Supreme Court’s decision in Eastern Book Company v D.B. Modak[25] is foundational to any analysis of AI authorship under Indian copyright law. The Court held that mere mechanical reproduction of text — however laborious — could not satisfy the originality requirement; the author must make genuine creative choices. Applied to AI-generated fashion, the principle is decisive: the question is not whether the output is original, but whether a legally cognisable human author made those creative choices. For AI-generated work, that author is absent, and the claim to copyright therefore fails at the threshold.
B. Thaler v Vidal: The DABUS Precedent
The most significant international precedent on AI authorship is Thaler v Vidal,[26] decided by the United States Court of Appeals for the Federal Circuit in 2022. Dr Stephen Thaler sought patent protection for inventions autonomously generated by his AI system, DABUS. The Court held that an “inventor” under the Patent Act must be a natural person, and that an AI system cannot be named as inventor regardless of how autonomously it generated the invention. Although arising in patent law, the reasoning applies equally to copyright and design statutes: statutory references to “person” consistently mean natural persons. Crucially, the Court declined to extend inventorship by judicial interpretation, reserving that step to the legislature — a posture Indian courts would do well to adopt.
C. Hermes International v Mason Rothschild: AI, NFTs, and Brand Integrity
In Hermes International v Mason Rothschild,[27] the Southern District of New York considered the sale of “MetaBirkins” — NFTs depicting AI-generated fur-covered versions of the iconic Hermès Birkin bag. The Court found in favour of Hermès, holding that the MetaBirkins were likely to cause consumer confusion as to source or sponsorship. The case confirms that AI-generated aesthetic works can infringe existing trade marks, and illustrates the tension between AI’s capacity for derivative creativity and the exclusivity functions of luxury trade marks. Importantly, it shows that trade mark law can be deployed defensively against AI-generated imitation — but it does not resolve the authorship question for AI-originated designs.
D. Christian Louboutin SAS v Abubaker and Ors: Design Protection in India
In Christian Louboutin SAS v Abubaker and Ors,[28] the Delhi High Court recognised that the iconic red lacquered sole could acquire trade mark distinctiveness through long and consistent commercial use, and granted an injunction against copycat products. The decision reflects the Indian judiciary’s increasing comfort with protecting the aesthetic identity of luxury fashion products under trade mark law. By extension, it raises a question for the AI era: could an AI-generated design element, deployed consistently over time, acquire registered trade mark protection in India? Technically yes — but only by attributing “origin” to a human or corporate applicant, which papers over rather than resolves the underlying machine-authorship problem.
V. Critical Analysis and Findings
The analysis reveals four key findings that highlight issues in the current legal framework for AI-generated fashion.
First, there is a significant gap in ownership. When an AI independently creates a valuable fashion design, no human made the creative decisions that copyright law aims to reward. Granting ownership to the developer, user, or deployer is more about policy choices than a true application of authorship principles. The judiciary is essentially improvising; a makeshift solution to a billion-dollar question is inadequate.[29]
Second, the registration-based system in the Designs Act 2000 offers a partial solution, but it lacks coherence. Allowing a human to register an AI-generated design rewards administrative action rather than creativity. This undermines the foundations of design protection and encourages a rush to register designs rather than to create.[30]
Third, the international framework is unclear and lacks guidance. TRIPS and the Berne Convention were not designed for creations by autonomous AI, and WIPO’s consultations have not resulted in binding standards. Therefore, India is compelled to protect creative works but cannot, under current law, determine ownership of AI-generated content.
Fourth, the unique commercial structure of the fashion industry exacerbates these issues. Fashion relies on the value of authorial identity. A garment’s value is tied to the story of who made it and why. When AI disrupts the human creative process, it destabilises copyright protection, design registration, and the narratives that support luxury pricing. An AI-created garment from a renowned brand raises serious questions about authenticity that brand management alone cannot resolve.[31]
The most promising yet underexplored option is to create a special protection system — a legal category for AI-generated works that does not claim human authorship while still providing protection. The UK’s computer-generated works provision in Section 9(3) of the Copyright, Designs and Patents Act 1988[32] gives authorship to the person who organises the work’s creation. While designed for a time before generative AI and imperfect in its application, it offers a legal basis. India should adopt and improve this approach by adding a requirement to disclose AI involvement, protecting consumers and maintaining integrity in authorship claims.
VI. Conclusion
The initial question — can a machine be a fashion designer legally — must be answered with a clear and regrettable no under the current framework of Indian IP law. The Copyright Act 1957, the Designs Act 2000, and the Trade Marks Act 1999 do not recognise a non-human creative agent as an author, designer, or rights-holder. This results in an ownership gap that will widen as AI systems become more capable and central to the fashion industry.
The solution is not to stretch current laws through creative judicial reasoning but to create clear and focused legislation. Three reform proposals emerge from this analysis.
First, India should amend the Copyright Act 1957 to include a specific provision for AI-generated works. Ownership should go to the human or corporate entity responsible for deploying the AI, along with a mandatory disclosure obligation to indicate the AI’s role in the creation.
Second, the Designs Act 2000 should be revised to require disclosure of AI participation in design creation. This change would prevent manipulation of the registration system to hide machine authorship and mislead consumers about design origins.
Third, India should lead efforts to establish a binding international protocol on AI authorship within the WIPO framework. This protocol should set minimum standards for disclosure, ownership, and duration of protection for AI-generated creative works across member states. In the fast-moving global fashion sector, where designs cross borders quickly and luxury brands operate worldwide, domestic reforms alone will not suffice.
The machine is now part of the design process. It sits at the sketch table, selects fabrics, and produces genuinely beautiful and commercially valuable designs. It is time for the law to take a front-row seat and engage earnestly with what it means to be a designer in the age of artificial intelligence.
Bibliography
Table of Cases
Christian Louboutin SAS v Abubaker and Ors CS(COMM) 714/2016 (Delhi High Court, 2018)
Eastern Book Company v D.B. Modak [2008] 1 SCC 1
Hermes International v Mason Rothschild, No. 22-cv-384 (S.D.N.Y. 2023)
Thaler v Vidal, 43 F.4th 1207 (Fed. Cir. 2022)
Yogendra Nath Naskar v Commissioner of Income Tax [1969] 1 SCC 555
Table of Legislation
Agreement on Trade-Related Aspects of Intellectual Property Rights (Marrakesh, 15 April 1994)
Berne Convention for the Protection of Literary and Artistic Works (Berne, 9 September 1886)
Copyright Act 1957
Copyright, Designs and Patents Act 1988 (UK)
Designs Act 2000
Trade Marks Act 1999
Secondary Sources
Abbott Ryan, ‘I Think, Therefore I Invent: Creative Computers and the Future of Patent Law’ (2016) 57 Boston College Law Review 1079
Cornish William, Intellectual Property: Patents, Copyright, Trade Marks and Allied Rights (9th edn, Sweet and Maxwell 2019)
Guadamuz Andres, ‘Artificial Intelligence and Copyright’ (2017) WIPO Magazine https://www.wipo.int/wipo_magazine/en/2017/05/article_0003.html accessed 2 June 2026
Locke John, Two Treatises of Government (first published 1689, Cambridge University Press 1988)
Scafidi Susan, ‘Intellectual Property and Fashion Design’ (2006) 1(1) Intellectual Property Law Review 115
Stitch Fix, ‘How We Use AI to Personalise Fashion’ (Stitch Fix, 2023) https://www.stitchfix.com/technology accessed 1 June 2026
World Intellectual Property Organization, ‘WIPO Conversation on Intellectual Property and Artificial Intelligence’ (WIPO, 2020) https://www.wipo.int/about-ip/en/artificial_intelligence/ accessed 3 June 2026
[1]See generally Stitch Fix, ‘How We Use AI to Personalise Fashion’ (Stitch Fix, 2023) https://www.stitchfix.com/technology accessed 1 June 2026.
[2]Ryan Abbott, ‘I Think, Therefore I Invent: Creative Computers and the Future of Patent Law’ (2016) 57 Boston College Law Review 1079, 1082.
[3]John Locke, Two Treatises of Government (first published 1689, Cambridge University Press 1988) 287–288.
[4]Copyright Act 1957, s 13.
[5]Designs Act 2000, s 2(d).
[6]Trade Marks Act 1999, s 2(1)(zb).
[7]Yogendra Nath Naskar v Commissioner of Income Tax [1969] 1 SCC 555, in which the Supreme Court of India recognised a Hindu idol as a juristic person capable of holding property.
[8]Andres Guadamuz, ‘Artificial Intelligence and Copyright’ (2017) WIPO Magazine https://www.wipo.int/wipo_magazine/en/2017/05/article_0003.html accessed 2 June 2026.
[9]Susan Scafidi, ‘Intellectual Property and Fashion Design’ (2006) 1(1) Intellectual Property Law Review 115, 117.
[10]Copyright Act 1957, s 13.
[11]Eastern Book Company v D.B. Modak [2008] 1 SCC 1.
[12]Copyright Act 1957, s 2(d).
[13]Copyright Act 1957, s 17.
[14]Copyright Act 1957, s 2(d)(vi).
[15]Designs Act 2000, s 2(d).
[16]Designs Act 2000, s 4.
[17]Designs Act 2000, s 6.
[18]Trade Marks Act 1999, s 2(1)(zb).
[19]Christian Louboutin SAS v Abubaker and Ors CS(COMM) 714/2016 (Delhi High Court, 2018).
[20]Trade Marks Act 1999, s 11(1).
[21]Agreement on Trade-Related Aspects of Intellectual Property Rights (Marrakesh, 15 April 1994) art 9.
[22]ibid art 10.
[23]Berne Convention for the Protection of Literary and Artistic Works (Berne, 9 September 1886) art 2.
[24]World Intellectual Property Organization, ‘WIPO Conversation on Intellectual Property and Artificial Intelligence’ (WIPO, 2020) https://www.wipo.int/about-ip/en/artificial_intelligence/ accessed 3 June 2026.
[25]Eastern Book Company v D.B. Modak [2008] 1 SCC 1.
[26]Thaler v Vidal, 43 F.4th 1207 (Fed. Cir. 2022).
[27]Hermes International v Mason Rothschild, No. 22-cv-384 (S.D.N.Y. 2023).
[28]Christian Louboutin SAS v Abubaker and Ors CS(COMM) 714/2016 (Delhi High Court, 2018).
[29]See Abbott (n 2) 1095–1096, criticising “fictional human author” approaches to AI-generated works.
[30]William Cornish, Intellectual Property: Patents, Copyright, Trade Marks and Allied Rights (9th edn, Sweet and Maxwell 2019) 651.
[31]Scafidi (n 9) 120–121.
[32]Copyright, Designs and Patents Act 1988 (UK), s 9(3), which vests authorship of computer-generated works in “the person by whom the arrangements necessary for the creation of the work are undertaken.”
