Home » Blog » AI-Generated Fashion Designs and Copyright: Who Owns the Runway of the Future this is my topic for a Blog.

AI-Generated Fashion Designs and Copyright: Who Owns the Runway of the Future this is my topic for a Blog.

Authored By: Manasvi Sharma

National Law Institute University, Bhopal

Abstract

The integration of generative AI into fashion design, producing original prints, patterns, and silhouettes, has created a significant legal vacuum within intellectual property law. This article examines the copyright implications of AI-generated fashion, addressing the central question of ownership: does copyright vest in the developer, the user, or does the work enter the public domain? Employing a comparative legal methodology, this research analyzes statutory frameworks and judicial interpretations across the United States, United Kingdom, and European Union.

The analysis reveals that existing copyright regimes, rooted in anthropocentric conceptions of authorship, are ill-equipped to address machine-generated creativity. The article finds that the absence of clear ownership rules creates commercial uncertainty for the fashion industry. It concludes by proposing potential pathways for legal reform, such as sui generis protections, to accommodate this emerging creative paradigm and provide necessary clarity for global luxury brands.

Introduction

The fashion industry, long defined by its reverence for human creativity and craftsmanship, stands at the threshold of a technological revolution. Artificial intelligence is no longer merely a tool for optimizing supply chains or personalizing customer recommendations, it has entered the creative process itself. Today, AI platforms can generate thousands of unique textile prints, propose novel garment silhouettes, and even conceptualize entire collections in moments, tasks that once demanded weeks of human labour and artistic intuition. This fusion of machine learning and design aesthetics promises unprecedented efficiency and creative possibility. Yet, beneath the surface of this innovation lies a profound legal disruption: the existing framework of copyright law, constructed around the romantic notion of the human author, is fundamentally unequipped to determine who, or what, owns the rights to these algorithmically generated creations.

This legal uncertainty strikes at the heart of the fashion industry’s commercial foundation. Copyright protection determines who can profit from a design, who can license it, who can prevent unauthorized copying, and who bears the risk of infringement. For emerging designers leveraging AI tools, for luxury houses investing in proprietary AI systems, and for technology developers licensing their platforms to the fashion sector, the absence of clear ownership rules creates significant commercial vulnerability. Without legal clarity, the runway of the future risks becoming a arena of litigation rather than innovation.

This article addresses a central and urgent question: under current copyright law, who owns an AI-generated fashion design? To answer this question, the article proceeds in five parts. Part II establishes the conceptual framework by examining the foundational principles of copyright, originality, authorship, and fixation, and their historical application to creative works. Part III undertakes a comparative legal analysis of how major jurisdictions, including the United States, the United Kingdom, and the European Union, have grappled with the question of AI authorship. Part IV examines relevant judicial interpretations and administrative decisions that illuminate the current legal landscape. Finally, Part V offers critical analysis and proposes potential pathways for legal reform to ensure that copyright law evolves in step with technological reality.

Background

To understand the legal predicament posed by AI-generated fashion designs, one must first examine the foundational principles upon which copyright protection rests. Copyright law, at its core, is designed to incentivise human creativity by granting creators exclusive rights over their original works. Three fundamental concepts are particularly relevant to the present inquiry: originality, authorship, and the idea/expression dichotomy.

The requirement of originality operates as the gatekeeper of copyright protection. In most jurisdictions, a work must originate from its author and demonstrate a minimal degree of creativity. The landmark decision in University of London Press v University Tutorial Press [1916] established that originality does not require novelty or inventive step, but rather that the work originated from the author and was not copied. This anthropocentric conception presupposes a human creator whose intellectual effort breathes life into the work. Similarly, the Berne Convention for the Protection of Literary and Artistic Works (1886), the cornerstone of international copyright law, implicitly assumes human authorship throughout its provisions, referring consistently to “authors” without contemplating non-human creators.

The concept of authorship carries distinct legal consequences. The author is presumptively the first owner of copyright and enjoys moral rights protections that safeguard their connection to the work. National copyright statutes reflect this human-centric orientation. The US Copyright Act defines a “work” as one “created by an author,” while the UK Copyright, Designs and Patents Act 1988, despite containing a provision for “computer-generated works” in Section 9(3), defines the author as “the person by whom the arrangements necessary for the creation of the work are undertaken.” The EU acquis, particularly the InfoSoc Directive (2001/29/EC), similarly operates within an anthropocentric framework, with the Court of Justice of the European Union emphasising that copyright protects the “author’s own intellectual creation.”

The idea/expression dichotomy further complicates the application of copyright to AI-generated fashion. Copyright protects the tangible expression of an idea, not the underlying idea itself. In fashion, this distinction proves particularly vexing: a garment’s silhouette or cut may be considered an unprotected idea, while its specific textile print or ornamental design may qualify as protectable expression. This doctrinal boundary, already contested in traditional fashion litigation, becomes increasingly unstable when AI systems generate designs that blur the line between conception and execution.

Prior scholarship has grappled with these tensions. Scholars such as Professors Jane Ginsburg and Andres Guadamuz have examined whether copyright frameworks can accommodate non-human creators, with most concluding that fundamental doctrinal reform would be required. The World Intellectual Property Organization has likewise initiated discussions on intellectual property and artificial intelligence, acknowledging that existing norms face unprecedented challenges from generative technologies.

Legal Analysis

The preceding section established that copyright law rests upon anthropocentric foundations. This part undertakes a critical examination of how three major jurisdictions—the United States, the United Kingdom, and the European Union—have grappled with the application of these principles to AI-generated works, with particular attention to the fashion context.

  1. The United States: Human Authorship as Constitutional Mandate

The United States maintains the most stringent position regarding human authorship. The US Copyright Office has repeatedly and explicitly refused to register works generated entirely by artificial intelligence, grounding its position in judicial precedent and constitutional interpretation. The seminal statement appears in the Compendium of US Copyright Office Practices (Third Edition), which provides that the Office will register “an original work of authorship, provided that the work was created by a human being.” This position derives from Burrow-Giles Lithographic Co v Sarony (1884), where the Supreme Court defined an “author” as “he to whom anything owes its origin.”

The Copyright Office’s recent decisions have reinforced this stance. In the Zarya of the Dawn case (2023), the Office determined that images generated by the AI platform Midjourney were not protectable, though it permitted copyright protection for the human-authored text and selection and arrangement of images. More emphatically, in Thaler v Perlmutter (2023), the US District Court for the District of Columbia affirmed the Office’s refusal to register a work created entirely by an AI system, holding that “human authorship is an essential part of a valid copyright claim.”

For fashion designers utilising AI tools, this position yields a binary outcome: designs generated autonomously by AI, without meaningful human creative input, fall outside copyright protection entirely and immediately enter the public domain. However, the analysis becomes more nuanced where human and machine contributions intersect. If a designer uses AI to generate a textile print but then materially modifies the output, or exercises creative control over the selection and arrangement of AI-generated elements, the resulting work may attract protection to the extent of human authorship. The burden rests upon the claimant to disentangle human from machine contribution—a factual inquiry that courts are only beginning to confront.

  1. The United Kingdom: A Statutory Exception

The United Kingdom presents a markedly different approach. Section 9(3) of the Copyright, Designs and Patents Act 1988 provides that “in the case of a literary, dramatic, musical or artistic work which is computer-generated, the author shall be taken to be the person by whom the arrangements necessary for the creation of the work are undertaken.” This provision, unique among major common law jurisdictions, explicitly contemplates non-human creation and designates a human author for legal purposes.

The definition of “computer-generated” appears in Section 178: a work is computer-generated if it is “generated by computer in circumstances such that there is no human author of the work.” The provision therefore applies precisely where the US position would deny protection entirely. For fashion designers in the UK, this creates a statutory framework capable of accommodating AI-generated textile prints, garment designs, and ornamental features—provided the designer can establish that they undertook the “arrangements necessary” for the work’s creation.

However, significant interpretive questions remain. What quantum of activity constitutes “arrangements necessary”? Does entering prompts into an AI platform suffice, or must the designer train the model, curate training data, or exercise post-generation selection? The provision has attracted sparse judicial consideration, leaving designers and their legal advisors to navigate uncertainty. The UK Intellectual Property Office has acknowledged these ambiguities and initiated consultation processes, but legislative clarification remains pending.

  1. The European Union: Originality as the Author’s Own Intellectual Creation

The European Union occupies an intermediate position. European copyright law, harmonised through directives and interpreted by the Court of Justice of the European Union (CJEU), predicates protection upon the work being the “author’s own intellectual creation.” This standard, articulated in Infopaq International A/S v Danske Dagblades Forening (C-5/08) and refined in subsequent jurisprudence, requires that the work reflect the author’s personality and express their creative choices.

The CJEU has not directly addressed AI-generated works. However, the logic of Infopaq suggests that autonomously generated AI output cannot satisfy the “own intellectual creation” standard, as it lacks the personal imprint of a human author. Works created through human use of AI tools may qualify if the human exercise of creative choices—selecting inputs, making aesthetic decisions, modifying outputs—results in a work bearing the human author’s personal touch.

For the fashion industry, this analysis aligns broadly with the US position, though the conceptual framework differs. EU law emphasises the creative choices reflected in the work rather than the ontological status of the author. A designer who employs AI to generate variations upon a theme and selects those expressing her creative vision may claim authorship of the resulting compilation or of individually modified elements.

Case Law Discussion

The legal principles examined above find concrete expression in judicial decisions that illuminate how courts approach questions of authorship, originality, and technological creativity. This section examines three cases of particular relevance to AI-generated fashion designs.

  1. Thaler v Perlmutter (US District Court, 2023)

The most direct statement on AI authorship in the United States arises from Thaler v Perlmutter. Dr Stephen Thaler sought copyright registration for a work titled “A Recent Entrance to Paradise,” which he stated was “autonomously created by a computer algorithm.” The Copyright Office refused registration on the ground that the work lacked human authorship, and Thaler challenged the decision in federal court.

The District Court affirmed the Office’s position, holding that “human authorship is an essential part of a valid copyright claim.” Judge Howell reasoned that copyright law has never protected works generated by non-human entities, from animal paintings to divine messages, and that extending protection to autonomous AI would require legislative action, not judicial interpretation. The court emphasised that copyright’s constitutional purpose—to promote progress by securing rights for authors—presupposes human creators whose incentive to create serves the public good.

For fashion designers, Thaler establishes that fully autonomous AI-generated designs enjoy no copyright protection in the United States. A textile print generated by an AI platform without meaningful human intervention enters the public domain immediately, available for any competitor to copy. This creates significant commercial risk for designers who rely heavily on automated generation tools.

  1. Infopaq International A/S v Danske Dagblades Forening (CJEU, 2009)

Though not concerning AI directly, Infopaq articulates the originality standard governing EU copyright law. The case involved whether eleven-word extracts of newspaper articles attracted copyright protection. The CJEU held that copyright protects works that are the “author’s own intellectual creation,” requiring that the work reflect the author’s personality and express their creative choices.

The significance for AI-generated fashion lies in the standard’s application. A designer who uses AI to generate a print must demonstrate that the resulting work embodies her intellectual creation. If she inputs prompts, selects among outputs, and modifies the chosen design, the work may satisfy Infopaq. However, if the AI operates autonomously with minimal human input, the resulting work likely falls outside protection. The CJEU’s emphasis on human creative choice aligns EU law with the anthropocentric assumptions underpinning international copyright.

  1. Nova Productions Ltd v Mazooma Games Ltd (England and Wales Court of Appeal, 2007)

This UK decision, though predating contemporary AI debates, provides guidance on computer-generated works under Section 9(3) of the CDPA 1988. The claimant alleged copyright infringement in a pool video game. The court considered whether screenshots generated by the game software during play attracted copyright protection.

The Court of Appeal held that the player did not qualify as author of the generated images, as the player’s input consisted of executing predetermined commands within the game’s framework. However, the court noted that the person “by whom the arrangements necessary for the creation of the work are undertaken” could be the programmer or game designer, depending on the circumstances. This reasoning suggests that for AI-generated fashion designs, the designer who operates the AI platform may qualify as author under Section 9(3) if they undertake the necessary arrangements—though the precise boundaries remain undefined.

The case confirms that UK law, unlike US law, possesses a statutory mechanism capable of accommodating non-human creation. For fashion designers, this offers potential protection for AI-generated works, provided they can establish their role in making the arrangements for creation.

Critical Analysis and Findings

The preceding analysis reveals a fundamental disconnect between technological reality and legal doctrine. Copyright law, constructed around the human creator, confronts a creative process that increasingly marginalises human contribution. This section offers critical commentary on the gaps exposed by this confrontation and evaluates the implications for the fashion industry.

  1. The Inadequacy of Binary Frameworks

The most striking finding is the inadequacy of existing legal frameworks to address the spectrum of human-machine collaboration. United States law adopts a binary approach: either a human author exists, and copyright attaches, or the work is autonomously generated, and protection is denied entirely. This framework cannot accommodate the reality of contemporary design practice, where designers train models, curate inputs, iteratively refine prompts, and exercise aesthetic judgment in selecting outputs. The law demands that courts disentangle human from machine contribution—a task for which existing doctrines of joint authorship and derivative works provide imperfect tools.

The UK’s Section 9(3), while ostensibly more accommodating, suffers from its own ambiguities. What constitutes “arrangements necessary” remains undefined, leaving designers uncertain whether their activities suffice. The provision, drafted in 1988, contemplated computer-assisted rather than computer-generated creation and has not been revisited in light of contemporary generative AI. Legislative inertia has rendered the provision potentially obsolete.

  1. The Commercial Consequences for Fashion

For the fashion industry, this legal uncertainty carries concrete commercial consequences. Emerging designers who leverage AI tools to compete with established houses face the greatest vulnerability. A designer who generates a distinctive print using Midjourney may find that print copied by a fast-fashion competitor, yet lack standing to sue for infringement. The absence of protection disincentivises innovation and advantages those with resources to employ human designers or litigate ownership questions.

Luxury houses face different but equally significant risks. Brands that invest in proprietary AI systems and training data may find that their investments generate unprotectable output, eroding the intellectual property portfolios that underpin brand value. The trade dress protections discussed in Part II may partially mitigate this risk, but copyright remains essential for protecting individual designs against exact copying.

  1. The Policy Vacuum

Perhaps most concerning is the policy vacuum at the international level. The Berne Convention, TRIPS Agreement, and WIPO Copyright Treaties operate on assumptions of human authorship that AI renders obsolete. No consensus exists among member states regarding how to address machine-generated creativity, creating fragmentation that disadvantages global fashion houses operating across multiple jurisdictions.

  1. Key Findings

Three principal findings emerge from this analysis. First, no major jurisdiction currently provides clear, adequate protection for autonomously AI-generated fashion designs. Second, the protection available for human-AI collaborative works varies significantly across jurisdictions, creating uncertainty for international fashion commerce. Third, legislative intervention, rather than judicial interpretation, is required to resolve the fundamental doctrinal tensions exposed by generative AI.

Conclusion

This article set out to answer a deceptively simple question: who owns the copyright in an AI-generated fashion design? The answer, as the preceding analysis demonstrates, depends critically upon jurisdiction, the degree of human involvement, and the interpretative choices of courts confronting technology their governing statutes never anticipated.

In the United States, autonomous AI-generated designs enjoy no protection whatsoever, while human-modified works may attract copyright to the extent of human contribution. In the United Kingdom, Section 9(3) of the CDPA 1988 offers a statutory pathway to protection, though its ambiguities leave designers in a state of uncertainty. In the European Union, the Infopaq standard requires that works bear the author’s own intellectual creation, a threshold that autonomous AI output cannot satisfy.

The runway of the future will feature designs emerging from human minds, machine algorithms, and every point along the continuum between them. Copyright law, in its current form, is not equipped to navigate this reality. Legislative intervention is urgently required to provide clarity, incentivise innovation, and ensure that the fashion industry’s creative future rests upon secure legal foundations. Until then, the question of ownership remains unanswered, and the industry bears the risk. 

References and Bibliography

Primary Sources

Legislation

European Union
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 L167/10 (InfoSoc Directive)

Italy
Legge 22 aprile 1941, n. 633 Protezione del diritto d’autore e di altri diritti connessi al suo esercizio (Copyright Law, as amended)

United Kingdom
Copyright, Designs and Patents Act 1988, c 48

United States
Copyright Act of 1976, 17 USC §§ 101–810
Lanham (Trademark) Act, 15 USC §§ 1051–1141n (2018)

International Treaties and Conventions

  • Berne Convention for the Protection of Literary and Artistic Works (adopted 9 September 1886, entered into force 5 December 1887) as last revised at Paris on 24 July 1971, 1161 UNTS 3
  • Marrakesh Agreement Establishing the World Trade Organization (adopted 15 April 1994, entered into force 1 January 1995) 1867 UNTS 154, annex 1C (‘Agreement on Trade-Related Aspects of Intellectual Property Rights’ (TRIPS Agreement))
  • WIPO Copyright Treaty (adopted 20 December 1996, entered into force 6 March 2002) 2186 UNTS 121

Cases

European Union

Case C-5/08 Infopaq International A/S v Danske Dagblades Forening [2009] ECR I-6569, [2009] ECDR 16

Case C-683/17 Cofemel – Sociedade de Vestuário SA v G-Star Raw CV [2019] ECR I-721, [2020] ECDR 1

United Kingdom

Donoghue v Stevenson [1932] AC 562 (HL)

Nova Productions Ltd v Mazooma Games Ltd [2007] EWCA Civ 219, [2007] EMLR 25

Response Clothing Ltd v The Edinburgh Woollen Mill Ltd [2020] EWHC 148 (IPEC), [2020] Bus LR 1562

University of London Press Ltd v University Tutorial Press Ltd [1916] 2 Ch 601 (Ch)

United States

Bartz v Anthropic PBC 787 F Supp 3d 1007 (ND Cal 2025)

Burrow-Giles Lithographic Co v Sarony 111 US 53 (1884)

Christian Louboutin SA v Yves Saint Laurent America Holding Inc 696 F 3d 206 (2d Cir 2012)

Kadrey v Meta Platforms Inc 788 F Supp 3d 1026 (ND Cal 2025)

Star Athletica LLC v Varsity Brands Inc 580 US 405 (2017)

Thaler v Perlmutter No 22-cv-1564 (DDC 18 August 2023) (WL 5333230)

Thomson Reuters Enterprise Centre GmbH v Ross Intelligence Inc 765 F Supp 3d 382 (D Del 2025)

Wal-Mart Stores Inc v Samara Brothers Inc 529 US 205 (2000)

Denmark

Ganni A/S v Steve Madden Ltd No BS-6112/2020-SHR (Maritime and Commercial High Court, 15 April 2021)

France

Maison Christian Dior v Evora Tribunal de grande instance Paris, 3e ch, 21 May 2013

Germany

Bundesgerichtshof [BGH] 20 February 2025, I ZR 16/24 (Birkenstock sandal designs)

Italy

Tribunale di Milano, Sezione Specializzata in Materia di Impresa, 25 January 2021, n 491/21 (Moon Boot)

Secondary Sources

Books

Stone D ‘Using Design Law to Protect Inventive Fashion in the UK and EU’ in Rosati E and Calboli I (eds) The Handbook of Fashion Law (Routledge 2025)

Yılmaztekin HK Artificial Intelligence, Design Law and Fashion (Routledge 2023)

Journal Articles

Ginsburg J ‘The Concept of Authorship in Comparative Copyright Law’ (2003) 52 DePaul Law Review 1063

Niyompatama MJ and Lapatoura I ‘Generative AI in fashion design creation: a copyright analysis of AI-assisted designs’ (2025) 20(10) Journal of Intellectual Property Law & Practice 654, https://doi.org/10.1093/jiplp/jpaf045

Raustiala K and Sprigman C ‘The Piracy Paradox: Innovation and Intellectual Property in Fashion Design’ (2006) 92 Virginia Law Review 1687

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

Yılmaztekin HK ‘Authorship of artificial intelligence: Global solutions and disjunctions’ in Artificial Intelligence, Design Law and Fashion (Routledge 2023) 127

Online Sources

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top