FROM RUNAWAY TO ALGORITHM: WHO OWNS AI- GENERATED FASHION DESIGNS?

ABSTRACT-

The expanding growth and use of Artificial Intelligence (AI) in the fashion industry has raised serious and complex legal questions regarding the ownership of AI- generated fashion designs. This article examines whether existing copyright frameworks, particularly under Indian law, are capable of addressing issues of authorship, originality and protection when creative outputs are produced through AI systems. The research further analyses competing ownership models, including the user-as-author, developer-as-author, and fashion brand-as-author approaches, while evaluating the relevance of the Indian Copyright Act, 1957, and comparative international frameworks. It further discusses significant judicial decisions through case laws to assess the continuing emphasis on human creativity in copyright jurisprudence. This study overall finds that current legal frameworks remain largely human-centric and inadequately equipped to regulate autonomous AI- generated fashion designs, thus, highlighting the need for legislative clarification and a balanced ownership framework that promotes innovations while also preserving the fundamental principles of copyright law.

KEYWORDS

Artificial Intelligence (AI), Fashion Law, Copyright, Authorship, Intellectual Property Rights, AI-Generated Designs, Fashion Industry, Ownership Rights.

1. INTRODUCTION

The Fashion Industry in its very form has constantly evolved alongside technological innovations. From mere manual sketching and pattern drafting to computer-aided design systems, technological advancements have significantly altered the manner in which fashion products have been conceived and produce[1] for years. In recent years, however, with the advent of Artificial Intelligence (AI), certain AI systems are now capable of producing original garment concepts, textile prints, color combinations, and even fashion collections within seconds.[2] Through machine learning algorithms, these systems have this ability of closely resembling the work which was traditionally performed by human designers[3].

The increasing adoption of AI in the field of fashion is becoming extremely common nowadays. Fashion houses and retailers are employing AI to predict consumer trends, automate certain design processes, personalize products, and reduce the production costs[4]. Simultaneously, law has struggled to determine how ownership rights should be allocated when creative works are generated through artificial intelligence.

Copyright laws have historically been built upon the very assumption that the creative works originate from human intellect and labor. Fundamental concepts such as originality, creativeness, original fashion textiles, authorship are deep rooted in the notion of human agency. AI-generated fashion designs challenge these foundational assumptions by introducing certain creative outputs that may involve varying degrees of human participation. Consequently, as a result, the legal procedures have often failed to address whether copyright ownership should be vested in the individual providing prompts, the developer who created the AI-system, the fashion company that commissioned the work, or whether such designs should remain unprotected under copyright law altogether. These are the questions that are frequently being asked.

The clear absence of legal guidance has become very problematic as certain AI- generated fashion designs, innovations, and textiles are gaining commercial significance. As a consequence, these very inconsistencies often create uncertainty for designers, technology developers, and many a times, fashion enterprises seek to exploit these AI- generated creations[5].
To address these core issues, this legal research article aims at examining certain conceptual foundations of copyright and AI- generated creativity and analyses the competing ownership models and several relevant legal principles, followed by a discussion of significant judicial decisions.
Finally, it evaluates shortcomings of the current legal frameworks and proposes certain fundamental reforms aimed at balancing innovation, commercial certainty, and the fundamental objectives of copyright law.

2. BACKGROUND / CONCEPTUAL FRAMEWORK

The origin of Artificial Intelligence (AI) in creative and fashion industries has resulted in fundamental challenge to the traditional understandings of authorship and ownership under the copyright law. Unlike conventional design software, AI- generated fashion designs in the domain of garments, textile patterns, accessories, or digital fashion concepts, generate creative content by identifying patterns within existing works and synthesizing them into new designs, thereby blurring the delicate distinction between creator and tool.

The Copyright Law has traditionally been founded upon the very principles of originality, creativity, and human authorship. Under the Indian Copyright Act, 1957, copyright subsists in original protected works, literary works, music works, or any other protected subject matter. Section 2(d) of the Act defines the “author” of a work, while Section 13 outlines the categories of works eligible for copyright protection[6]. However, the legislation was enacted long before the development of autonomous AI systems and does not explicably address ownership of works generated through artificial intelligence. In the United States, copyright protection is generally contingent upon human authorship, a principle which is repeatedly affirmed by the courts and US Copyright Office[7]. At the very contrast, U.S. Copyright Office. In contrast, the United Kingdom adopts a comparatively flexible approach through Section 9(3) of the Copyright, Designs and Patents Act, 1988, which provides that in case of computer-generated works, authorship may be vested upon the person who undertakes and carries the responsibility of making the arrangements necessary for the creation of work.[8]

Several academic scholarships remain dividend on the issue of AI- generated creativity. While some scholars view that AI should be used merely as a sophisticated tool, akin to traditional design software, with ownership fledged upon the human user exercising creative control, others contend that AI systems challenge the very conceptual foundation of copyright law and that they should require a development of new legal frameworks[9].

Hence, the dispute surrounding AI-generated fashion designs raise a fundamental legal question: whether the current copyright doctrines and laws, which are rooted in human creativity, can effectively regulate ownership in an era where the algorithms increasingly participate in the creative process or not. But this very uncertainty forms the conceptual foundation for analysing competing ownership models and evaluating the adequacy for contemporary copyright law.

3. LEGAL ANALYSIS

At the heart and core of copyright laws, there lies the very principle that- protection is granted to original works resulting from human intellectual effort. Copyright systems in this case have historically been founded on the assumption that “creativity” itself is an inherently human attribute. In India, the Copyright Act, 1957 recognizes authorship as a prerequisite for copyright ownership and links protection to several identifiable creators[10]. Similarly, on the other hand, American copyright jurisprudence has consistently emphasized human authorship as a fundamental requirement for protection[11].

The advent of generative AI as discussed earlier challenges the foundational assumption. As a result, the traditional connection between the author and work becomes increasingly difficult to establish. If copyright exists to reward human form creativity, it often becomes uncertain that whether outputs generated substantially by algorithms can satisfy this requirement.[12]

The inadequacy and the failure to take into account the current evolving issues becomes particularly evident in fashion design, where creativity often emerges as a result through collaborative interaction between human prompts and machine-generated outputs. The current copyright frameworks provide little guidance on determining the very thresholds of human involvement necessary for authorship, creating uncertainty regarding ownership and enforcement.[13]

3.1 Competing Models of Copyright Ownership

3.1a The User-As-Author Model

One approach suggests that copyright should be bestowed on the individuals who operates the AI system. Under this model, the user contributes to developing creative inputs through prompts, instructions, selection of outputs, and subsequent modifications. The argument is that AI functions merely as a sophisticated tool, analogues to graphic design software or digital illustration platforms.[14]

However, this approach raises certain practical concerns as well. Modern AI systems can produce highly detailed fashion designs from relatively simple prompts such as “generate a sustainable haute couture evening gown inspired by nature.” If such minimal input can create such fashion designs, the originality requirement risks become diluted. Copyright protection could be granted despite only limited human creative contribution, thereby weakening and disrupting the traditional justification for copyright.[15]

3.1b The Developer-as-Author Model

A second approach proposes that ownership should be bestowed on the developers who design and train the AI- system. Proponents often argue that developers should create the technological architecture that enables the generation of creative outputs and by doing it, they make a substantial contribution the final result.[16]

Such position is very difficult to justify under the name of established copyright principles. Although, developers create the underlying software, they do not exercise the direct creative control over industrial fashion designs produced by the users. By granting ownership to developers, it could lead to excessive concentration of intellectual property rights in the hands of technology companies.[17]

 3.1c The Fashion Brand-as-Author Model

If we look from a commercial standpoint, fashion brands often invest huge number of resources in acquiring AI tools, employing prompt engineers, and commercializing AI- generated collections. Consequently, as a result, some argue that ownership should be bestowed on the commissioning entity in a manner similar to employer ownership under work-for-hire arrangements.[18]

While this approach promotes commercial certainty, it conflicts with the core principle of copyright law. Extending ownership solely on the basis of investment risks transforms copyright from a creativity-based right into a commercial monopoly.

3.2 Should Ai-Generated Fashion Designs Receive Copyright Protection at All?

Can an artwork created entirely by Artificial Intelligence get copyright protection under Indian law? This question has come up as one of the most disputed questions in the field of AI and fashion. A more radical position argues that fully autonomous AI- generated fashion designs should not get copyright protection. Since there’s no human author that exists in such cases, the work would enter the public domain immediately upon creation. This approach often aligns with the traditional view that copyright exists to incentivize human form of creativity[19] rather than machine-generated production and designs.

Nevertheless, denying protection entirely may create undesirable consequences and circumstances. Given the speed with which the fashion trends are being replicated and commercialized, the absence of legal protection could result in undermining of incentives required for technological advancement within the industry.[20]

In Thaler v. Perlmutter (2023), involving Stephan L. Thaler’s AI system DABUS, the U.S. court reaffirmed that copyright protection requires human authorship, denying protection to a work created autonomously by AI[21]. In India, Section 2(d)(vi) of the Copyright Act, 1957 recognizes the “person who causes the work to be created” as the author of computer-generated works. Therefore, the applicability of AI- generated fashion remains uncertain and creates ambiguity regarding ownership of AI- generated fashion designs[22].

3.3 The Need for a Reconceptualised Legal Framework

The kind of legal uncertainty surrounding the AI- generated fashion deigns demonstrates that existing copyright frameworks are becoming increasingly inadequate at this point. Neither the Indian Copyright Act, 1957, nor most contemporary copyright statutes were being drafted with the most autonomous AI systems in their mind. As AI assumes a far greater role in creative decision making, the traditional concepts of originality and authorship are becoming increasingly strained.

A more balanced and concise approach would recognize copyright only where a certain meaningful degree of human creative control is exercised over the design process. In domains where AI operates autonomously and independently, with minimal contribution, a separate sui generis regime may provide a more appropriate solution. Thus, when we talk about fashion design, where innovation, speed, and originality are critical economic assets, certain legal reforms have become increasingly necessary in these domains to ensure both certainty and fairness in the allocation of ownership rights.

4. CASE LAW DISCUSSION

The dispute in respect to the ownership in AI- generated fashion designs remain largely unresolved, debated, particularly in India where no court has yet directly addressed the issue. The following cases illustrate how courts have approached these principles and their implications for AI- generated fashion work.

• Thaler v. Perlmutter (2023)

In Thaler v. Perlmutter, Stephan L. Thaler sought copyright protection for an artwork generated autonomously by his AI system, DABUS.  The U.S. Copyright Office rejected this application on the grounds that copyright protection needs human authorship. The District Court upheld the decision by holding that copyright law protects works originating from human based intellectual efforts and does not extend to creations generated solely by machines.[23]

The legal principle established that human authorship remains a fundamental requirement for copyright protection. Although not binding in India, the decision itself in highly persuasive because Indian copyright law is very similar on the concept of an “author” under Section 2(d) of the Copyright Act, 1957[24]. Consequently, if a fashion design is generated entirely by AI without meaningful human involvement, Indian courts may adopt a similar approach and deny copyright protections too.

•  Eastern Book Company v. D.B. Modak (2008)

In Eastern Book Company v. D.B. Modak, the Supreme Court of India examined whether edited judicial reports were sufficiently original to qualify for copyright protection. Here, in this case, the Court rejected the traditional “sweat of the brow” doctrine and upheld that copyright requires a minimum degree of creativity and intellectual contribution.[25]

This judgement is particularly relevant to certain AI- generated fashion designs because it establishes originality as the mere cornerstone of copyright protection in India. Accordingly, the decision sparks of the suggestion that copyright protection should depend largely upon the extent of human creative input rather than the mere operation of an AI tool or software.

• G. Anand v. Deluxe Films (1978)

In R.G. Anand v. Deluxe Films, the Supreme Court emphasized that copyright protects the original expression of an idea rather than the idea itself.[26] The principles that were used in this case serves as the direct implications for AI- generated fashion designs. If originality arises primarily from a designer’s prompts, modifications, and creative choices, authorship should arguably be bestowed upon the human creator. However, where the AI system independently determines the final design, identifying a legally recognizable author or a creator becomes significantly more challenging. The ambiguity itself exposes a gap within the current Indian copyright framework.

Therefore, in a nutshell, these all cases collectively demonstrate that both Indian and international copyright jurisprudence continue to prioritize human creativity, originality, and intellectual contribution. All their combined influence suggests that Indian courts are likely to favour human- centric approach to ownership, while also exposing and clarifying the urgent need for legislative clarification regarding AI- generated fashion designs[27].

5. CRITICAL ANALYSIS / FINDINGS

It is often reflected that the rapid integration of Artificial Intelligence into fashion design has exposed significant gaps in the already existing intellectual property frameworks. Including India, where copyright law in most jurisdictions is fundamentally premised on human creativity and authorship. Consequently, as a result, AI- generated fashion designs challenge the very traditional legal assumptions in regards to ownership, originality, and accountability. While the AI systems can independently work on generating novel patterns, garments, and design concepts, the law in its very form remains uncertain as to whether such outputs qualify for copyright protection and, if so, who should be recognised as the author.[28]

In the case, Thaler v. Perlmutter, the decision of the United States Copyright Office reaffirmed that copyright protection requires human creative input[29]. Similarly, the Court of Justice of the European Union has repeatedly emphasized that a protected work must reflect the author’s “own intellectual creation”[30]. These developments indicate that existing copyright regimes continue to prioritise human agency over machine- generated creativity. Nevertheless, such an approach may inadequately address the realities of some increasingly autonomous AI systems used within the fashion industry.

It may be observed that Indian copyright law is comparatively more adaptable. Section 2(d)(vi) of the Copyright Act, 1957 recognizes the person who causes a computer-generated work to be created as its author[31]. It is submitted that Section 2(d)(vi) may no longer be sufficient to regulate modern generative AI systems.

However, provisions were enacted long before the emergence of generative AI and does not clarify itself whether authorship belongs to the programmer, the AI developer, the fashion house, or the user who inputs prompts. Accordingly, the mere absence of judicial interpretation creates a substantial uncertainty for designers and businesses who seek to commercialize AI- generated fashion outputs.

Furthermore, with the expansion of more policy implications, there was a rise of ownership disputes. Comparative departments particularly the European Union’s AI regulatory framework and ongoing discussions within the World Intellectual Property Organisation (WIPO), suggest a growing preference for a more balanced governance rather than full legal recognition of AI as an author[32]. Therefore, a more nuanced and civil legislative approach is necessary. It is proposed that copyright protection should remain human- centric while also introducing statutory guidelines to determine ownership where AI substantially contributes to the creative process. And, such reforms would promote innovations, enhance legal certainty, and better align intellectual property law with the technological realities in the fashion industry.

6. CONCLUSION

The emergence of Artificial Intelligence in the fashion industry has fundamentally challenged the conventional concepts of authorship and ownership within the Intellectual Property (IP) law. This very research aims at finding the existing copyright frameworks, both in India and internationally, that remained largely centred on human creativity and are therefore ill- equipped to address the complexities of AI- generated fashion designs[33]. While AI can independently produce innovative and commercially valuable designs, current legal regimes generally refuse to recognise AI as an author, thereby creating uncertainty regarding ownership and protection.

Accordingly based on the research, the answer to- who owns AI- generated fashion designs- remains largely unresolved and disputed topic. Under the present framework, ownership is most likely to be attributed to the human actor involved in the creative process, whether the user, developer, or commissioning entity. However, the absence of a clear statutory guidance continues to generate ambiguity and potential disputes.

Ultimately, as fashion increasingly transitions from the designer’s sketchbook to the algorithm’s output, the central challenge for intellectual property law will be, to balance technological innovation with the enduring legal principle that creativity deserves recognition and protection. The journey from runway to algorithm therefore marks not the end of authorship, but the beginning of a new legal conversation about its meaning in the digital age.[34]

REFERENCE(S):

1. Pamela Stecker, The Fashion Design Manual: Principles, Practice and Techniques (Macmillan 2012) 12–18.
2. World Intellectual Property Organization, ‘Generative Artificial Intelligence and Intellectual Property: An Overview for Policymakers’ (WIPO 2024) 5–8.
3. World Intellectual Property Organization, WIPO Technology Trends 2019: Artificial Intelligence (WIPO 2019) 19–24.
4. McKinsey & Company, ‘The State of Fashion 2025’ (McKinsey 2025) 38–41.
5. European Union Intellectual Property Office, Study on Artificial Intelligence and Intellectual Property Rights (EUIPO 2022) 45–52.
6. Copyright Act 1957, ss 2(d), 13.
7. Thaler v Perlmutter 687 F Supp 3d 140 (DDC 2023).
8. Copyright, Designs and Patents Act 1988 (UK), s 9(3).
9. Jane C Ginsburg and Luke Ali Budiardjo, ‘Authors and Machines’ (2019) 34 Berkeley Technology Law Journal 343, 350–359.
10. Jane C Ginsburg and Luke Ali Budiardjo, ‘Authors and Machines’ (2019) 34 Berkeley Technology Law Journal 343, 347–355.
11. World Intellectual Property Organization, Generative Artificial Intelligence and Intellectual Property: An Overview for Policymakers (WIPO 2024) 14–19.
12. Ryan Abbott, The Reasonable Robot: Artificial Intelligence and the Law (Cambridge University Press 2020) 74–79.
13. Jane C Ginsburg and Luke Ali Budiardjo, ‘Authors and Machines’ (2019) 34 Berkeley Technology Law Journal 343, 356–358.
14. Annemarie Bridy, ‘Coding Creativity: Copyright and the Artificially Intelligent Author’ (2012) 5 Stanford Technology Law Review 1, 15–18.
15. University of London Press Ltd v University Tutorial Press Ltd [1916] 2 Ch 601.
16. World Intellectual Property Organization, Generative Artificial Intelligence and     Intellectual Property: An Overview for Policymakers (WIPO 2024) 21–23.
17. Eastern Book Company v D.B. Modak (2008) 1 SCC 1.
18. G. Anand v Deluxe Films (1978) 4 SCC 11
19. Jane C Ginsburg and Luke Ali Budiardjo, ‘Authors and Machines’ (2019) 34 Berkeley Technology Law Journal 343, 356–359.
20. Jane C Ginsburg and Luke Ali Budiardjo, ‘Authors and Machines’ (2019) 34 Berkeley Technology Law Journal 343, 347–359.
21. Thaler v Perlmutter 687 F Supp 3d 140 (DDC 2023).
22. Infopaq International A/S v Danske Dagblades Forening (C-5/08) [2009] ECR I-6569.

[1] Pamela Stecker, The Fashion Design Manual: Principles, Practice and Techniques (Macmillan 2012) 12–18.

[2] World Intellectual Property Organization, ‘Generative Artificial Intelligence and Intellectual Property: An Overview for Policymakers’ (WIPO 2024) 5–8.

[3] World Intellectual Property Organization, WIPO Technology Trends 2019: Artificial Intelligence (WIPO 2019) 19–24.

[4] McKinsey & Company, ‘The State of Fashion 2025’ (McKinsey 2025) 38–41.

[5] European Union Intellectual Property Office, Study on Artificial Intelligence and Intellectual Property Rights (EUIPO 2022) 45–52.

[6] Copyright Act 1957, ss 2(d), 13.

[7] Thaler v Perlmutter 687 F Supp 3d 140 (DDC 2023).

[8] Copyright, Designs and Patents Act 1988 (UK), s 9(3).

[9] Jane C Ginsburg and Luke Ali Budiardjo, ‘Authors and Machines’ (2019) 34 Berkeley Technology Law Journal 343, 350–359.

[10] Copyright Act 1957, ss 2(d), 13.

[11] Thaler v Perlmutter 687 F Supp 3d 140 (DDC 2023).

[12] Jane C Ginsburg and Luke Ali Budiardjo, ‘Authors and Machines’ (2019) 34 Berkeley Technology Law Journal 343, 347–355.

[13] World Intellectual Property Organization, Generative Artificial Intelligence and Intellectual Property: An Overview for Policymakers (WIPO 2024) 14–19.

[14] Jane C Ginsburg and Luke Ali Budiardjo, ‘Authors and Machines’ (2019) 34 Berkeley Technology Law Journal 343, 350–353.

[15] World Intellectual Property Organization, Generative Artificial Intelligence and Intellectual Property: An Overview for Policymakers (WIPO 2024) 16–18.

[16] Ryan Abbott, The Reasonable Robot: Artificial Intelligence and the Law (Cambridge University Press 2020) 74–79.

[17] Jane C Ginsburg and Luke Ali Budiardjo, ‘Authors and Machines’ (2019) 34 Berkeley Technology Law Journal 343, 356–358.

[18] Annemarie Bridy, ‘Coding Creativity: Copyright and the Artificially Intelligent Author’ (2012) 5 Stanford Technology Law Review 1, 15–18.

[19] University of London Press Ltd v University Tutorial Press Ltd [1916] 2 Ch 601.

[20] World Intellectual Property Organization, Generative Artificial Intelligence and Intellectual Property: An Overview for Policymakers (WIPO 2024) 21–23.

[21] Thaler v Perlmutter 687 F Supp 3d 140 (DDC 2023).

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

[23] Thaler v Perlmutter 687 F Supp 3d 140 (DDC 2023).

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

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

[26] R.G. Anand v Deluxe Films (1978) 4 SCC 118.

[27] Jane C Ginsburg and Luke Ali Budiardjo, ‘Authors and Machines’ (2019) 34 Berkeley Technology Law Journal 343, 356–359.

[28] Jane C Ginsburg and Luke Ali Budiardjo, ‘Authors and Machines’ (2019) 34 Berkeley Technology Law Journal 343, 347–359.

[29] Thaler v Perlmutter 687 F Supp 3d 140 (DDC 2023).

[30] Infopaq International A/S v Danske Dagblades Forening (C-5/08) [2009] ECR I-6569.

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

[32] World Intellectual Property Organization, Generative Artificial Intelligence and Intellectual Property: An Overview for Policymakers (WIPO 2024) 18–25.

[33] Thaler v Perlmutter 687 F Supp 3d 140 (DDC 2023).

[34] World Intellectual Property Organization, Generative Artificial Intelligence and Intellectual Property: An Overview for Policymakers (WIPO 2024) 22–25.