The Legal Status Of AI-Generated Fashion In Authorship, Ownership And Design Rights

Author: LawSphereX

Category: Property & Real Estate

Published: 2026-08-23T21:25:16.593047+00:00

Updated: 2026-08-26T17:05:34.474+00:00

THE LEGAL STATUS OF AI-GENERATED FASHION IN

AUTHORSHIP, OWNERSHIP AND DESIGN RIGHTS

Mohan Kumar S.K., LL.M. (Intellectual Property and Trade Law), School of Law, CHRIST (Deemed to be University), Bengaluru, India

ABSTRACT

This study analyses the emerging legal issues concerning AI-created fashion designs, in particular with respect to authorship, ownership, originality, and design protection under Indian and International intellectual property (IP) laws. It examines academic and policy sources setting out how copyright and design laws might operate when artificial intelligence is responsible for the creation of clothing and accessories. The review includes academic works analyzing AI-generated creativity, government reports, and legal commentary. Issues considered include the need for a human author, the level of originality and whether design right regimes will protect or fail to protect AI-assisted fashion designs. Indian points of view are highlighted with comparison to U. S. , EU and WIPO views.

The survey points to the fact that most of the present laws still presuppose human creativity such as the Indian Copyright Act attributes authorship only to a person who “causes” a computer to make a work. At large there is consensus that AI works are not covered by copyright at this time and therefore legislation is needed to fill the gap. Significantly, we observe voids in material in India, Indian laws have yet to explicitly deal with generative AI, design law is not considered in a large number of studies. Through drawing a parallel with recent legal developments worldwide, like UK’s “computer-generated works” rule, EU’s data-mining folklores, and US courts discussing AI art, the review shows the gaps in Indian AI authorship, ownership, and design rights clarification. It ends with highlighting open issues e. g.

, who owns the design created by an AI-engine and a strong plea for legal adaptation. This analysis provides guidance for further research on modifying IP law for AI-generated fashion innovation. Keywords: AI-generated fashion, Authorship, Design protection, Originality, Intellectual Property law.

INTRODUCTION

Artificial Intelligence (AI) is one of the wonders of the 21st century that is massively changing the fashion sector. It has started to rebrand its creativity and design through fully or partially self-driven machines. By the use of DALL·E, Midjourney, and CLO3D, not only is the fashion industry able to come up with original pieces of art that could also be marketable, but the fashion industry can also extend the limits of its artistic expression while at the same time challenging the set legal norms1. The merging of technology and fashion has softened the distinction between human and machine creativity; thus, certitude is highly demanded in terms of originality, authorship, and ownership of the works. AI-generated fashion features massive potential to bring new ideas, reduce costs, and make the customer experience one of a kind.

Nevertheless, this new form of creation also unsettles the traditional legal system, which has always been based on the concept of human authorship and originality. Unclear legal status of AI-driven results leads to a variety of consequences for human designers, companies, and government officials, as the issue of protection and control is yet to be solved. This article is a thorough commentary on the legal position of artificial intelligencegenerated fashion with special reference to the issues of authorship, ownership, and design rights. Through a review of current academic works and international policies, it lays out the insufficiencies in the existing legislative framework and speculates on the possible ways to ensure further reforms that could keep up with such developments in the fashion industry. 2

1. Conceptual Framework of AI and Fashion

The connection between Artificial Intelligence (AI) and fashion continues to develop at a fast pace as it transforms both design processes and business structures. The heritage of fashion as a human-centred creative space now merges with algorithmic processes to create new artistic possibilities. The merging of AI with fashion technology creates problems that stretch past both creative aspects and technological issues toward intellectual property rights and ownership. The purpose of this section is to establish a conceptual basis which explains AI applications in

1 OpenAI, ‘DALL·E 2 Overview’ (OpenAI, 2022) https://openai.com/dall-e-2 accessed 18 September 2025;

MidJourney, ‘User Guide’ (MidJourney, 2023) https://docs.midjourney.com accessed 18 September 2025; CLO Virtual Fashion, ‘CLO3D Overview’ (CLO3D, 2023) https://www.clo3d.com accessed 18 September 2025. 2 WIPO, ‘Artificial Intelligence and Intellectual Property Policy’ (WIPO, 2020) https://www.wipo.int/aboutip/en/artificial_intelligence/ accessed 18 September 2025; Andres Guadamuz, ‘Do Androids Dream of Electric Copyright? Comparative Analysis of Originality in AI Works’ (2017) 2(1) Intellectual Property Quarterly 1.

fashion design. This chapter examines AI's part in creative development before describing its uses in fashion production then explores artistic authorship controversies and finally analyses property rights disputes with their legal consequences.

1.1 Artificial Intelligence in Creative Production

Modern computing systems that mimic human mental processes through learning, reasoning and problem-solving fall under the current definition of AI. Modern machine learning algorithms together with deep neural networks have moved AI beyond automation tasks to creative domains. Generative AI algorithms absorb huge volumes of images and text data which enables them to produce fresh outputs that imitate human creative outputs. The fashion industry benefits from DALL·E (OpenAI) and MidJourney and Stable Diffusion which use textual inputs to generate distinctive visual content that expands design possibilities. 3 The fashion industry benefits from two specific software programs named CLO3D and Browzwear which let designers build computer-generated realistic prototypes of clothing.

4 AI technology cuts down prototyping expenses alongside development duration and raw material consumption compared to conventional design approaches. The speed and scalability of AI-powered creativity are transforming the fashion design process. Modern AI systems have the ability to generate multiple design variations at a rate that surpasses human creativity since designers typically create limited sketches in a day but AI generates many options in minutes. 5 The ability of machines to generate design concepts challenges our traditional understanding of creativity as an exclusively human capacity by raising questions about whether artificial intelligence can genuinely contribute to design.

1.2 AI in Fashion: From Design to Virtual Runways

The use of AI in the fashion industry has been crossing the boundary of just design to all sorts of applications that include production, marketing, retail, and consumer engagement besides fashion design. The creativity of design in the AI algorithms come up with designs, colour 3 OpenAI, ‘DALL·E 2 Overview’ (2022) https://openai.com/dall-e-2 accessed 18 September 2025; MidJourney, ‘User Guide’ (2023) https://docs.midjourney.com accessed 18 September 2025; Stability AI, ‘Stable Diffusion’ (2023) https://stability.ai accessed 18 September 2025.

4 CLO Virtual Fashion, ‘CLO3D Overview’ (2023) https://www.clo3d.com accessed 18 September 2025;

Browzwear, ‘About Us’ (2023) https://browzwear.com accessed 18 September 2025. 5 Andres Guadamuz, ‘Artificial Intelligence and Copyright’ (2017) 2 Intellectual Property Quarterly 169.

options, and styles that can go along with the trend that has been predicted. For example, DALL·E and MidJourney may create fashion visualizations that are then used as a starting point for designing actual clothing. 6 Further, Virtual Sampling and Prototyping; CLO3D is a program that lets designers see clothes on digital people, showing the fabric and how it flows and moves. 7 This helps the supply chain reduce waste because fewer physical samples are needed. The Trend Prediction, AI systems may study consumer tastes, social media activity, and historical sales data to predict next fashion trends. In the Virtual Fashion and Metaverse, the creation of AI is trendy clothing is increasingly the sale as digital assets (e. g. , NFTs) for avatars in virtual spaces, which, is bridging the gap between physical and digital fashion.

8 Finally, in the Eco-Friendliness, through resource optimization and the elimination of waste per the AI textile, fashion may become more wondrously sustainable as compared to the past. These are the instances where AI technology is no longer just an instrument for-mindedness but a stakeholder in the greater fashion business that is hereby affecting how fashion is produced, consumed, and experienced.

1.3 Authorship and Creativity in AI-Generated Fashion

One of the most debated controversies is the eligibility and the extent of A. I in being recognised as authors. The human intellect is the basis for creativity in the tradition and the copyright and design frameworks work on it. The creative work is associated with individual authorship, skill, and creativity. 9 In the case of AI-generated fashion, the problem worsens. Imagine an AI that creates a dress design entirely on its own and all it has is the data it was given to learn from. It is fascinating and challenging to determine the limits of human creativity and the onset of machine creativity. To solve these three scholars, propose models. In AI as A Tool, this model interprets AI as a paintbrush or a sewing machine. The originality and the creative intent rest with a human using the system. Consider a designer who gives instructions to Midjourney.

That person continues to be the author of the design. Then, AI as a Co-Creator, AI co-creation with a human collaborator and human creativity involve sharing authorship or attributing it to the human who led the entire process. Additionally, AI as an Independent Author, this is the most radical view, which treats AI as an independent creator. 6 WIPO, ‘Artificial Intelligence and Intellectual Property’ (WIPO, 2020) https://www. wipo. int/aboutip/en/artificial_intelligence/ accessed 18 September 2025.

7 CLO Virtual Fashion (n 2).

8 Hermès International v Mason Rothschild No 22-cv-384 (JSR), 2023 WL 1458126 (SDNY 2023).

9 E Book Co v D B Modak (2008) 1 SCC 1 (SC, India).

While this is intriguing from a perspective of philosophy, the idea faces several legal issues as most of the existing intellectual property laws.10 In the fashion industry, this discussion stands out the most as the opinion of various designers is shaped by the originality of a design and the branding of a company. With significant inputs from AI, who is responsible for the value of the end product the designer of the product or the machine?

1.4 Ownership and Control of AI-Generated Output

As we consider ‘authorship,’ the more pressing concern relates to ‘ownership. ’ Traditional copyright frameworks grant ownership to the author, barring any assignment, or a creation that emerges within the bounds of employment. 11 Fashion generated by AI, on the other hand, presents a blurred version of ownership. Assuming a designer employs AI software following a license agreement, ownership may align with the provisions of that license. A number of AI platforms (for example, DALL·E and MidJourney) declare that users hold rights to the outputs they generate and that the AI system and the training data used are still the developer’s property. 12 Imagine AI is the sole creator of the designs without any considerable human guidance. In that case, the main concern is whether such designs can be treated as “works” under copyright or design law.

If the answer is negative, they stand a chance of becoming public. This in turn restrains creators from leveraging their designs commercially. In partnership models, conflicts are likely to emerge involving the designer, the AI’s developer, and perhaps the owners of the training data that was used in the AI’s dataset. This poses a great threat towards the fashion industry because the industry thrives on brand identity and exclusivity.

1.5 Implications for Intellectual Property Law

The Conceptual debates have significant implications for intellectual property (IP) frameworks such as the Copyright Law, Usually, it safeguards the original artistic works that have been fixed in some kind of medium. One of the biggest problems is the necessity to prove that the author is human and that it was original when AI had a significant role in it.13 Design Law, it 10 Anna Ronkainen, ‘AI Authors and the Copyright Law’ (2020) 42(3) European Intellectual Property Review

11 The Copyright Act 1957, s 17 (India).

12 OpenAI, ‘Terms of Use’ (2023) https://openai.com/policies/terms-of-use accessed 18 September 2025; MidJourney, ‘Terms of Service’ (2023)

13 Amarnath Sehgal v Union of India 2005 (30) PTC 253 (Del HC).

covers just the new, beautiful features of the industrial products. Nevertheless, if a design generated by AI is not directly linked to a human, then it is questioned whether it is possible to classify it as 'design' in terms of legal definitions or not. In the Patent Law, a singular instance can be an AI to come up with new ways for fabric production or material science. The discussion about AI inventorship, e. g. in the DABUS global scenarios, is very much applicable here. 14 Trademark and Personality Rights, the AI-generated avatars dressed in digital clothing are raising branding and personality rights issues, especially in the event that they very much resemble real celebrities. These differences point to the fact that presently the IP regimes are not adequate, as they take into account human authorship and not machine-generated creativity.

Additionally, in the case law of Star Athletica, L. L. C. v. Varsity Brands, Inc. , the U. S. Supreme Court decided that the artistic aspects of cheerleading uniforms were copyrightable if they could be separated from their utilitarian function. AI-generated fashion shares the same concept, where the digital designs can be considered as the original creative works before the product is made by the industry. 15 Further, fashion uniqueness in India has also been acknowledged by the court. In Christian Louboutin SAS v. Nakul Bajaj, the Delhi High Court, among other things, reiterated the criteria for trademark identity and recognizability, hence confirming the distinct red sole mark as a 'famous mark' and emphasizing the judicial idea of giving strong protection to the unique features of fashion. 16

1.6 Philosophical and Ethical Considerations

Aside from legal and business aspects, the burgeoning of AI in fashion field has a wide range of questions that includes philosophy and ethics: What is creativity? Theoretically, if creativity is seen as one of the qualities that consist in coming up with new and useful ideas, AI may be considered as creative entities. Nevertheless, if creativity implies self-awareness or purpose, then AI is lacking. Authenticity of the culture, fashion is art as well as a product of culture and society. AI-assisted designs with the aid of worldwide datasets will result in culturally untrue or homogenized outputs. The displacement of workers, the process of automating design has the effect of pushing out the junior side of the fashion industry, skills such as drawing, and craftsmanship, thus the fear of economic injustice in the industry that might occur.

14 Thaler v Comptroller-General of Patents, Designs and Trade Marks [2021] EWCA Civ 1374; Thaler v Vidal

43 F.4th 1207 (Fed Cir 2022); European Patent Office, J 8/20 (DABUS), Decision of 21 December 2021.

15Star Athletica, LLC v Varsity Brands, Inc 580 US 405 (2017).

16 Christian Louboutin SAS v Nakul Bajaj 2018 (253) DLT 728 (Del HC).

The broader debate is whether non-human entities can be recognized as rights-holders. In Amarnath Sehgal v. Union of India (2005) 30 PTC 253 (Del), the Delhi High Court emphasized the importance of moral rights, highlighting the author’s personal connection to their work. Moral rights presuppose human authorship and cannot logically extend to AI. 17 The Naruto v. Slater (2018) “monkey selfie case” in the United States is particularly relevant. The court held that animals cannot claim copyright, underscoring the principle that authorship requires legal personality. This analogy has been repeatedly invoked in AI debates, suggesting that AI, like animals, cannot own copyright. 18 This section has laid out the conceptual framework of AI-generated fashion, specifying the enormous change capacity of the new technology, and the difficulties that arise in the legal and the creative ecosystems.

It has been shown that AI is no longer a mere assistant but a main actor in fashion design, virtual retail, and digital identity. The issue of intellectual property law has become complicated due to the indistinguishable authorship and ownership rights. The arguments of this section are like the prerequisites for the following part, which focuses on the Indian legal framework. It investigates the Copyright Act, 1957 and the Designs Act, 2000 concerning originality, authorship, and the protection of AI-generated fashion, as well as if the existing laws are adequate in handling the interruptions caused by AI.

2. The Indian Legal Framework Governing AI-Generated Fashion

2.1 Copyright Act, 1957 and AI-Generated Fashion

The Copyright Act, 1957 is the cornerstone of intellectual property protection for creative works in India. Fashion designs, sketches, textile patterns, and digital renderings generated by AI must be assessed under its provision.19

2.1.1 Originality and Creativity Requirement

Section 13(1)(a) is the provision that safeguards "original artistic works" those might be fashion sketches or digital patterns created by AI. Nevertheless, the main problem is the originality test.20 The Supreme Court in Eastern Book Company v. D.B. Modak (2008) 1 SCC 1 has made

17 Amarnath Sehgal v. Union of India (2005) 30 PTC 253 (Del)

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

19 The Copyright Act 1957 (India).

20 Copyright Act 1957, s 13(1)(a).

it clear that originality requires a minimum degree of creativity, thus it has discarded the "sweat of the brow" approach which was the earlier stance. Taking this logic to an AI output fully generated by algorithms and thus there should be no significant human intervention, the court might decide that such a creation is not an "original artistic work".21 On the other hand, if a designer is the one who makes a detailed prompt, selects, and edits, in that case, such activities could meet the requirement of creativity. For example, a designer who applied CLO3D to make a digital visualization of a saree with innovative draping could be the one to first claim that it is the result of their mental contribution to the design that makes it original.

2.1.2 Authorship and Ownership

Section 2(d) of the Act defines the “author” of an artistic work as the artist who produces the work22. Non-human entities do not receive consideration from the statute. The copyright law does not grant authorship status to AI systems including DALL·E and MidJourney. The author becomes the initial copyright owner according to Section 17 which states the author receives ownership rights first except for prior agreements and employment contracts.23 The human user or commissioning party who uses AI represents the most likely candidate to obtain ownership rights. The UK court determined in Nova Productions Ltd. v. Mazooma Games Ltd. [2007] EWCA Civ 219 that individuals who organize the essential steps for developing computergenerated works qualify as authors.24 The Indian legal system does not have an equivalent provision but courts might apply this interpretation to AI situations.

2.1.3 Moral Rights

Section 57 establishes author moral rights that protect both attribution rights and defense against work distortions.25 The 2005 Delhi High Court ruling in Amarnath Sehgal v. Union of India declared moral rights as untransferable human creations. The personal relationship between author and work required by moral rights makes it impossible for AI to receive these rights.26 The principle that AI lacks independent authorship receives confirmation through this concept.

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

22 Copyright Act 1957, s 2(d).

23 Copyright Act 1957, s 17.

24 Nova Productions Ltd v Mazooma Games Ltd [2007] EWCA Civ 219.

25 Copyright Act 1957, s 57.

26 Amarnath Sehgal v Union of India 2005 (30) PTC 253 (Del HC).

2.2 Designs Act, 2000 and Fashion Application

The Designs Act, 2000, protects the visual features of shape, configuration, pattern, ornamentation, or composition applied to an article.27 AI-generated garments, once converted into physical products, often fall under this regime.

2.3 Artistic Work vs. Design Dichotomy

The Copyright Act Section 2(d) removes designs from protection when these designs become eligible for registration under the Designs Act through industrial application.28 The Delhi High Court explained through its Microfibres Inc. v. Girdhar & Co. (2009) 4 SCC 396 decision that artistic works applied to fifty or more articles become subject to the Designs Act. AIgenerated fashion operates under this established principle.29 The protection moves from copyright to design law when an AI produces digital embroidery patterns that get massproduced on garments.

2.4 Protection and Term

The registration of designs under Section 11 provides protection for ten years with an option to renew for five years.30 The requirements for protection under design law differ from copyright requirements since it demands both novelty and distinctiveness but not strict originality. AI-generated patterns which avoid direct copies of existing works might meet the registration criteria. The Designs Act of India is increasingly being used to safeguard textile patterns which includes Sabyasachi’s Banarasi motifs31. AI-generated motifs that maintain originality could register for design protections under current regulations.

2.5 Trademark Protection in Fashion and AI

The legal framework of trademarks interacts directly with AI-generated fashion products particularly when branding digital or physical clothing lines. Protection under the Trade Marks Act 1999 covers unique logos and names as well as colours and shapes linked to fashion houses. The Delhi High Court recognized the red sole as a trademark in Christian Louboutin SAS v.

27 The Designs Act 2000 (India).

28 Copyright Act 1957, s 2(d).

29Microfibres Inc v Girdhar & Co 2009 (40) PTC 519 (Del HC).

30 Designs Act 2000, s 11.

31 See ‘Sabyasachi Couture: Protection of Textile Designs under the Designs Act’ (Bar and Bench, 2020) https://www.barandbench.com accessed 18 September 2025.

Nakul Bajaj (2018) which marked a new direction for fashion IP protection. The designercreated AI logo or brand name can receive registration status when it achieves distinctiveness32. AI systems generating outputs that match existing brand elements may trigger future disputes which will evaluate potential infringement violations.

2.6 Overlap with Information Technology Act, 2000

The IT Act 2000 lacks explicit provisions about authorship but its electronic record and liability regulations apply to the matter. The use of unauthorized datasets containing copyrighted textile designs by AI tools could trigger Section 66C (identity theft) and Section 66E (violation of privacy) for enforcement.33 The liability protection of intermediaries under Section 79 exists when they prove they took proper precautions and did not have awareness of infringement activities.34 The legal system offers a method to handle disagreements between fashion copyright infringements and AI-generated outputs.

2.7 Judicial Approaches to Fashion IP and AI: India and Comparative Perspectives

Indian courts have progressively developed principles that, although not yet extended to AIgenerated fashion, offer direction on originality, authorship, moral rights, and the intersection of copyright, design, and trademarks. In Eastern Book Company v D B Modak, the Supreme Court of India clarified the standard of originality, holding that there must be a "modicum of creativity" in order to be granted copyright protection, thus abandoning the previous "sweat of the brow" doctrine. 35 This maxim does have immediate relevance to how courts may evaluate originality in AI-designed works: if products are solely machine-created and lack meaningful human input, they may not meet this minimum, but human-influenced inputs possibly might.

In Microfibres Inc v Girdhar & Co, the Delhi High Court ruled that protection then moves to the Designs Act once an artistic work is used for over fifty articles, a principle directly applicable to AI work, such as embroidery patterns or motifs, which may begin life as artistic work but lose copyright when being mass-produced36. Parallelly, in Amarnath Sehgal v Union of India, the Court reaffirmed moral rights under Section 57 of the Copyright Act, emphasizing

32 Christian Louboutin SAS v Nakul Bajaj 2018 (253) DLT 728 (Del HC).

33Information Technology Act 2000, ss 66C, 66E. 34Information Technology Act 2000, s 79. 35Eastern Book Company v D B Modak (2008) 1 SCC 1 (SC, India).

36 Microfibers Inc. vs Girdhar & Co. & Anr. on 28 May, 2009

the author's unprotectable relationship with the work, which AI cannot achieve because of its non-personhood. 37 In Christian Louboutin SAS v Nakul Bajaj, robust trademark protection was granted to the iconic red sole, evidencing judicial willingness to protect unique fashion details, a rule that could extend to AI-generated logos or design elements. 38 Comparative cases resonate these tendencies: in Star Athletica, LLC v Varsity Brands, Inc. , the US Supreme Court determined that artistic elements that are apart from utilitarian function are eligible for copyright protection, an argument that can be applied to AI-drawn sketches;39 and in Hermès International v Mason Rothschild, the court expanded trademark protection into cyberspace by declaring "Meta Birkin" NFTs infringed Hermès' trade dress.

40 Lastly, the DABUS cases in the UK, US, and EU disapproved of the acknowledgement of AI as an author or inventor, confirming that only natural persons are eligible, thus reinforcing the worldwide judicial practice favoring human-oriented intellectual property while reserving the existing doctrines to safeguard originality, design uniqueness, and brand identity within the fashion industry. 41

2.8 Challenges and Research Gap

Despite statutory coverage under the Copyright Act 1957, Designs Act 2000, and Information Technology Act 2000, critical gaps persist in regulating AI-generated fashion. Indian law does not recognise non-human authorship, and Section 2(d) of the Copyright Act explicitly limits authorship to human creators, leaving AI outputs without clarity on protection. The copyright– design overlap further complicates matters, since artistic works like digital embroidery or motifs, once applied industrially, lose copyright protection and shift into the design regime, creating uncertainty for AI-driven creations. Ownership too is unclear, with a majority of AI platforms like DALL·E and MidJourney granting rights via terms of service in favour of developers, leaving designers at the mercy of contractual terms instead of statutory protection.

Liability is also unclear: Section 79 of the IT Act does provide a safe-harbour to intermediaries, but does not touch upon dataset misuse in training of AI or define responsibility where offending designs are created. These challenges collectively expose a research gap, as Indian

37 Amarnath Sehgal v. Union of India, 2005 (30) P.T.C. 253 (Del. HC) (India).

38 Christian Louboutin SAS v. Nakul Bajaj, 2018 (253) D.L.T. 728 (Del. HC) (India).

39 Star Athletica, L.L.C. v. Varsity Brands, Inc., 580 U.S. 405 (2017).

40 Hermès Int’l v. Rothschild, No. 22-cv-384 (JSR), 2023 WL 1458126 (S.D.N.Y. Feb. 8, 2023).

41Thaler v. Comptroller-General of Patents, Designs and Trade Marks, [2021] EWCA (Civ) 1374 (UK); Thaler v. Vidal, 43 F.4th 1207 (Fed. Cir. 2022) (U.S.); European Patent Office, J 8/20 (DABUS), Decision of Dec. 21, 2021.

IP law remains firmly human-centric and insufficiently equipped to address authorship, ownership, liability, and the copyright–design interface in the age of AI-generated fashion.42 43

44 45.

Copyright Act, Designs Act, and Trade Marks Act. Originality, industrial application, and distinctiveness are the norms of judicial interpretation that will be useful in adaptation. The lack of legislative acknowledgement of AI as a creative force, though, leaves much uncertainty. Until reform is legislated, the balance will fall to human intervention, with the courts likely treating designers, curators, and commissioning parties as authors and owners. However, while AI continues to reshape the creative process, legislative change is inevitable in order to ensure that fashion innovation is suitably guarded in India.

3. Comparative Analysis on the US, UK and EU

The global fashion world functions across various legal systems, all trying to come to terms with the impact of AI-generated work. A comparative study of the United States, the United Kingdom, and the European Union offers an insight into how intellectual property regimes meet the challenges posed by authorship, ownership, and design rights under AI fashion.

3.1 United States: Human Authorship and Expansion into Virtual Fashion.

Intellectual property law in the United States, which is the main part of Title 17 of the United States Code, is very protective of copyrights that are formed originally and creatively and are preserved in a physical medium. 46 The Copyright Office has always declared that an author is a person and has to be creative; thus, it has acknowledged no registration for works produced totally by AI, like in the 'Zarya of the Dawn' case regarding AI-generated comic art. 47 In Star Athletica, LLC v. Varsity Brands, Inc. , the Supreme Court found that the artistic features of cheerleading uniforms were copyrightable if they could be conceptually separated from their utilitarian function, an idea which can also be applied to AI-generated sketches and digital 42Copyright Act, No. 14 of 1957, INDIA CODE (1957), section 2(d). 43Designs Act, No. 16 of 2000, INDIA CODE (2000). 44See, e. g.

, OpenAI, Terms of Use, https://openai. com/policies/terms-of-use (last visited Sept. 18, 2025); MidJourney, Terms of Service, https://docs. midjourney. com/docs/terms-of-service (last visited Sept. 18, 2025).

45 Information Technology Act, No. 21 of 2000, INDIA CODE (2000), section 79.

4617 U.S.C. section 102(a) (2018).

47U.S. Copyright Office, Zarya of the Dawn (Feb. 21, 2023), https://copyright.gov/rulings/zarya-of-the-dawn (last visited Sept. 18, 2025).

patterns.48 Trademark law has undergone a similar change: the court in Hermès International v. Rothschild decided that the “MetaBirkin” NFTs were committing trade dress infringement of Hermès’ handbag, and thus brand protection was possible in the virtual market.49 Together, these cases show that while the US system still requires human creativity for authorship, it is also capable of offering rights to AI-assisted works if there is evidence of human creative intervention.

3.2 United Kingdom: Computer-Generated Works and Narrow Judicial Reading.

Section 9(3) of the UK Copyright, Designs and Patents Act 1988 is a notable provision that only mentions 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. "50 This legislative novelty opens the door to the possibility of AI works being recognized, but the interpretation by the courts is very limited. In the Court of Appeal decision in Nova Productions Ltd v. Mazooma Games Ltd, the court confined the notion of author to those who had control over the software and not to the software itself. 51 The UK Supreme Court in the DABUS case did not accept the AI inventorship; thus, it confirmed that the definition of the author and the inventor still refers to human beings only.

52 Hence, the courts in the UK, while they have been theoretically acknowledging the computer-generated works as per the law, have actually been very cautious and have not gone far in attributing authorship to non-human entities.

3.3 European Union: Originality and the “Author's Own Intellectual Creation”

Standard. Harmonization of the EU copyright legal decisions is achieved through directives and the Court of Justice of the European Union (CJEU). In the case Infopaq International A/S v. Danske Dagblades Forening, the CJEU decided that originality implies the work to be the “author’s own intellectual creation. ”53 This criterion, as defined, suggests human authorship; thus, AIgenerated works that do not have a meaningful human contribution are automatically out of the 48Star Athletica, L. L. C. v. Varsity Brands, Inc. , 580 U. S. 405 (2017). 49Hermès Int’l v. Rothschild, No. 22-cv-384 (JSR), 2023 WL 1458126 (S. D. N. Y. Feb. 8, 2023). 50Copyright, Designs and Patents Act 1988, c. 48, § 9(3) (UK). 51Nova Prods. Ltd. v. Mazooma Games Ltd. , [2007] EWCA (Civ) 219 (UK). 52Thaler v. Comptroller-General of Patents, Designs & Trade Marks, [2023] UKSC 49. 53Infopaq Int’l A/S v.

Danske Dagblades Forening, Case C-5/08, EU:C:2009:465.

scope of this classification. In the area of design, Regulation (EC) No 6/2002 establishes the rights for "Community Designs," which must be new and have an individual character; thus, if the human creative contribution is clearly visible, it may be possible that the protection is extended to AI-assisted fashion. The trademark law in the EU is strong and the protection is not limited to traditional marks but also includes non-traditional marks such as colors and patterns which is the case that can be seen in Louboutin v. Van Haren.54 However, similar to the situations in the US and UK, the EU law has also turned down the acknowledgement of AI as a creator or an inventor, as can be seen in the European Patent Office's decision to reject the DABUS applications.55

3.4 Comparative Observation

A consistent pattern can be seen in different legal areas: that copyright, design, and trademark

laws are able to support AI-assisted fashion, but the creation of the work is still considered to be the human’s responsibility. In the US, the main way is through court decisions, which set the limit for cases of non-human authorship; in the UK, there is a provision in the law for works generated by computers, but the interpretation is rather strict, and the EU applies a harmonized test of “intellectual creation” to decide authorship. The systems each show a gradual change in IP law to accommodate the new technology, but they have not come to a clear answer on the matter of fully autonomous AI authorship yet.

4. Conceptual Framework of Celebrity Rights in the AI Era

AI-generated fashion raises not only issues of copyright and design rights but also the broader question of celebrity rights, where AI can replicate the likeness, identity, and brand value of public figures.

4.1 Personality and Publicity Rights in Fashion

One of the key aspects of personality rights is to prevent unauthorized usage of an individual's name, image, likeness, and unique characteristics for commercial purposes without the consent of such individual. The judges of India, in the case of ICC Development v. Arvee Enterprises, confirmed the concept of publicity rights, stating that the celebrity's personality is a separate

54 Christian Louboutin v. Van Haren Schoenen BV, Case C-163/16, EU:C:2018:423.

55 European Patent Office, J 8/20 (DABUS), Decision of Dec. 21, 2021.

source of commercial value.56 Artificial Intelligence in the fashion industry can replicate "digital twins" of famous people wearing AI-generated clothing, which can cause a violation of the celebrity's rights.

4.2 Trademark and Passing Off Protection

The fashion brands of celebrities normally depend on the registration of trademarks of their names, logos, and signatures. In DM Entertainment vs. Baby Gift House, the Delhi High Court stopped the illegal use of singer Daler Mehndi's image on toys, releasing the court order treating the act as misrepresentation and passing off.57 Correspondingly, in the United States, the court case White v. Samsung Electronics America, Inc. resulted in a decision that the portrayal of a robot similar to Vanna White violated her right of publicity.58 These are the cases that point to the danger of AI-generated avatars that can impersonate celebrity identities without authorization in the virtual fashion market.

4.3 Copyright Performers’ Rights

According to the Copyright Act 1957, Sections 38 and 38A give performers’ rights that protect against the unauthorized reproduction of performances.59 One such infringement could be the use of AI to create avatars that wear virtual clothes that are inspired by the performances of celebrities (for instance, digital recreations of stage outfits) in order to use them in the AI fashion domain. Nevertheless, as AI is not a performer, only the aspect of human likeness shall be considered.

4.4 Data Protection and Privacy Concern

One of the methods that AI uses to create systems is by collecting images of celebrities from social networks, which causes issues with the privacy of the users' data according to regulations like GDPR of the EU and the Digital Personal Data Protection Act of India 2023.60 In addition, the use of biometric data, facial images, or body scans for the purpose of creating virtual fashion driven by AI may implicate the violation of the user's privacy rights, along with the

56 ICC Dev. (Int’l) Ltd. v. Arvee Enters., 2003 (26) P.T.C. 245 (Del. HC) (India).

57 DM Ent. (P) Ltd. v. Baby Gift House, 2010 (42) P.T.C. 520 (Del. HC) (India).

58White v. Samsung Elecs. Am., Inc., 971 F.2d 1395 (9th Cir. 1992). 59Copyright Act, No. 14 of 1957, INDIA CODE (1957), §§ 38, 38A. 60 Regulation (EU) 2016/679 (General Data Protection Regulation), art. 4; Digital Personal Data Protection Act, No. 22 of 2023, INDIA CODE (2023).

infringement of intellectual property rights.

4.5 Implication for AI-Generated Fashion

The merging of the fashion and celebrity civilizations has just emphasized how vital it is that personality rights be known in addition to the intellectual property rights. The use of artificial intelligence to copy the image of a famous person has the potential not merely to lower the value of the brand but also to trigger basic rights related to human dignity, autonomy, and giving consent. Careful legislation must provide that celebrities, especially hyped in AI fashion as a result of a commercial venture without their authorization should be the first to have control over their identity.

5. Critical Analysis and Research Finding

5.1 Critical Analysis

The Indian fashion legal framework that governs the fashion industry is still framed in a way to prioritise humans and therefore presents certain challenges for works created by AI. The Copyright Law of India identifies “author” only as a natural person;61 the Supreme Court in Eastern Book Company v. D. B. Modak has gone further by stating that originality presumes human creative power, thus filtering out creations that are non-human with no or little human contribution. 62 Besides confusing issues of copyright–design overlap induce also mixed feelings as under Section 15 of the Copyright Act, artistic works cease to have copyright once they are applied industrially to more than fifty articles63, a point that the court in Microfibres Inc. v. Girdhar & Co. has confirmed most clearly64, i. e.

AI outputs like digital motifs might be initially copyrightable and just after they are mass-produced to switch abruptly to the design law. Section 57 of moral rights also demonstrates the laws' human bias: the connection between the author and the work was at the centre of the Delhi High Court’s decision in Amarnath Sehgal v. Union of India, thus effectively AI is not considered in the recognition. 65 However, trademark law seems to be more flexible in adapting, as demonstrated by the case of Christian Louboutin SAS v. Nakul Bajaj, where the famous red sole was considered a 61Copyright Act, No. 14 of 1957, INDIA CODE (1957), section 2(d). 62E. Book Co. v. D. B. Modak, (2008) 1 S. C. C. 1 (India). 63Copyright Act, No. 14 of 1957, INDIA CODE (1957), section 15. 64Microfibres Inc. v. Girdhar & Co. , (2009) 4 S. C. C. 396 (India). 65Amarnath Sehgal v. Union of India, 2005 (30) P. T. C.

253 (Del. HC) (India).

trademark66 and the US case of Hermès International v. Rothschild, allowing virtual “MetaBirkin” NFTs to get protection. 67 The instances mentioned above show that although copyright and design legislation are unable to handle AI, trademarks can still reconcile with the new technologies, provided that the main focus is on the uniqueness and brand identity. Still, the question of who owns what and who is accountable for it remains open with AI platforms such as DALL·E and MidJourney, which grant rights via terms of service that favour creators, leaving users dependent on contracts rather than law for protection.

68 The IT Act, despite offering intermediaries a safe harbour under Section 7969, is silent on dataset misuse and accountability for AI developers, whereas the EU’s GDPR and Digital Services Act set out more defined responsibilities for data handling and platform liability70, thus providing a roadmap for Indian reforms.

5.2 Research Findings

This study underlines the major issues in three points: the authorship concept is very much a human one and is therefore under Indian law; AI-generated fashion that is not copyrighted faces the issue of the protection of its moral rights; the overlaps between the copyright and the design sectors allow instability to arise, especially from the point of view of those works that by starting as artistic sketches become industrially reproduced, thus losing copyright as per Section 15 of the Copyright Act;71 and the gaps in terms of contracts and liabilities put the designers in a risky position since ownership and accountability depend mainly on private agreements and the few safe harbours under the IT Act. 72 On the other hand, US law is different in a way that it clearly opposes AI authorship, but still modifies trademarks for the digital realm, as in Hermès International v. Rothschild.

73 UK law gives a statutory definition of computer-generated works under Section 9(3) of the CDPA and narrows its interpretation. 74 whereas EU law strictly requires “intellectual creation” and 66Christian Louboutin SAS v. Nakul Bajaj, 2018 (253) D. L. T. 728 (Del. HC) (India). 67Hermès Int’l v. Rothschild, No. 22-cv-384 (JSR), 2023 WL 1458126 (S. D. N. Y. Feb. 8, 2023). 68OpenAI, Terms of Use, https://openai. com/policies/terms-of-use (last visited Sept. 18, 2025); MidJourney, Terms of Service, https://docs. midjourney. com/docs/terms-of-service (last visited Sept. 18, 2025). 69Information Technology Act, No. 21 of 2000, INDIA CODE (2000), § 79. 70Regulation (EU) 2016/679 (General Data Protection Regulation); Regulation (EU) 2022/2065 (Digital Services Act). 71Copyright Act, No. 14 of 1957, INDIA CODE (1957), § 15. 72Information Technology Act, No. 21 of 2000, INDIA CODE (2000), § 79.

73Hermès Int’l v. Rothschild, No. 22-cv-384 (JSR), 2023 WL 1458126 (S. D. N. Y. Feb. 8, 2023). 74Copyright, Designs and Patents Act 1988, c. 48, § 9(3) (UK).

therefore, human input is implied, as in Infopaq International A/S v. Danske Dagblades Forening.75 These three are not perfect in solving the problem of AI authorship, but nevertheless, they all realize the necessity of reform. In India, a new route will be the implementation of more precise authorship rules, the integration of copyright and design safeguards, and the setting of liability norms for AI platforms. Without such steps, AIgenerated fashion will still be in a state of legal uncertainty, which will be a great obstacle to its development and protection.

6. Conclusion and Suggestion

AI-generated fashion analysis just shows that India and the rest of the world have made some changes in their intellectual property laws to accommodate the new reality, but these changes are not enough to address the peculiarities of machine creativity. Some changes are necessary to have legal security. In India, these changes should be towards a good balance between innovation and safety. The initial proposal is the creation of a legal definition of authorship in the Copyright Act, a clarification specifically dealing with AI-assisted works. It is possible to take inspiration from the UK Section 9(3) of the CDPA as a model, which credits authorship of computer-generated works to the person who has done the most work to the extent necessary.

If there were a comparable clause in India, it would guarantee that AI users who create the AI tools and give them the creative direction would be the authors recognised, thus, they would be the ones to fill the current unoccupied space. Secondly, the overlap between copyright and design should be revisited. The strict fifty-article rule in Section 15 of the Copyright Act practically removes the rights of fashion designs at the moment they reach their commercial value. A changed method could enable technology to provide the designs produced by AI with both art and industry, thus ensuring the originality is not compromised by the scalability of the works. On top of that, standards for liability in the Information Technology Act need to be upgraded.

The sources of the data for the training of AI platforms should be the basis on which AI platforms are held responsible, with a cover of a few specific measures that prohibit the utilization of infringing content. By incorporating statutory obligations similar to those under the EU’s Digital Services Act and GDPR, India would be in harmony with the world’s best 75Infopaq Int’l A/S v. Danske Dagblades Forening, Case C-5/08, EU:C:2009:465.

practices, thus giving creators more protection from the unauthorised exploitation of their works. Fourthly, the reliance on compliance with the AI developers' terms of service that are solely created by them, should be minimized through establishing a legal baseline for data ownership in AI outputs. Creators who put their technical expertise and guidance into AIenabled operations ought to get first ownership rights unless it is specifically indicated otherwise, thus making the power of negotiation between creators and platforms more equitable. Lastly, the rights of personality and publicity have to become more robust in order to avoid the occurrence of AI-powered impostors that simulate the likeness of celebrities in the fashion industry.

The social acknowledgement of publicity rights, when in harmony with trade mark and copyright, would protect not only the economic but also the dignitary aspects of individuals. To put it in a nutshell, research demonstrates in the end that the trademarks have relatively succeeded in their adaptation to the AI-driven fashion. However, the copyright and design laws in India are still lagging behind; they are too human-centric and can't handle the blurred authorship and ownership realities of AI. The verdicts both in India and foreign countries expose a trend that is common all over the world, which is the unwillingness to recognize AI as an author. At the same time, they also point to the readiness of courts to change the principles, taking into account new technological features if a human is involved in the creative process.

The main research result is that India has to carefully weigh the options: on the one hand, maintaining the doctrine of human authorship as the guiding principle of intellectual property law, and on the other, making explicit statutory provisions that acknowledge AI as a creative tool. The absence of reform will mean the legal uncertainty over AI-generated fashion, which will discourage both innovation and protection. The reform will turn India into a modern and adaptive IP regime that not only secures the interests of designers but also facilitates the technological progress and protects the uniqueness of fashion.