CTRL+Culture - Artificial Intelligence And The Exploitation Of Traditional Designs In Fashion
Author: LawSphereX
Category: Technology, Cyber & Privacy
Published:
Updated:
CTRL+CULTURE: ARTIFICIAL INTELLIGENCE AND THE
EXPLOITATION OF TRADITIONAL DESIGNS IN FASHION
Diya Dilip, BITS Law School, Mumbai
ABSTRACT
Traditional cultural expressions (TCEs) are intergenerational knowledge systems that are interwoven with the social, cultural and economic identities of traditional artisanal communities. There has been documentation of the vulnerability of TCEs in the context of misappropriation by global fashion houses, but with the integration of Artificial Intelligence, there poses a new threat. The central question is when AI and large fashion industries use TCEs as input to produce outputs, at what point does it shift from inspiration to infringement and why does the Indian framework fail to protect these vulnerable originating communities? Prior research submits that while current Intellectual Property laws provide partial protections, it has foundational gaps where collective rights disappear.
This paper addresses how these gaps interact with AI driven infringement using an analysis of Indian IP statutes, International models and theoretical engagement with Epistemic Injustice and Postcolonialism. It attempts to define the fine line between inspiration and infringement and the possible implications of such infringement that undermines cultural significance and rights. It is argued AI outputs derived from TCEs are rarely transformative and create a shadow market that is detrimental to traditional artisans. The paper advocates for a Sui Generis model that acknowledges communal authorship and rights, striving to harmonize the existence of AI and TCEs in the fashion industry.
The findings of this paper underscore the need to reconsider the relationship between technological innovation and long standing culture, prompting the question if technological progress can occur without erasing the very culture it draws from.
1. Introduction
India is known for its vibrant culture and diversity. With this culture comes centuries old expressions of art in various forms that have been passed down and remain the pride of communities. Across India, traditional artisans craft practices that sustain both their livelihoods and cultural identities and these practices and cultural knowledge are held collectively, embedded in social lives. Some of these craftsmanship are the result of long standing marginalization and reflect stories that go beyond the surface level understanding. However, with the evolution of AI and large fashion industries that use traditional motifs without any attribution to the original artists , the ecology of traditional craftsmanship has been disrupted. With the coming of AI that mimics or incorporates traditional work, it's not only economic exploitation at risk but also the erasure of culture.
With the emergence of AI and fashion, it's becoming clearer that the current protections available are not sufficient to address the needs of the hour. Copyright laws revolve around individual authorship and identifiable creative origin, whereas traditional craftsmanship and cultural expressions are communal and intergenerational. Designs are created by rural caste based communities, tribal groups or village artisans who lack knowledge of protections available and also rarely have access to said protections. This creates a disproportionate advantage to large fashion industries and AI allowing them to benefit and exploit vulnerable artisanal communities. It's is to be noted that though there is no widely accepted definition of traditional cultural expression, Article 31.
1 of the United Nations Declaration on the Rights of Indigenous Peoples (UNDRIP) 2007 states “Indigenous peoples have the rights to maintain, control, protect and develop their cultural heritage, traditional knowledge and traditional cultural expressions, as wells as the manifestations of their sciences, technologies and cultures, including human and genetic sources, seeds, medicines, knowledge of the properties of fauna and flora, oral traditions, literatures, designs, sports and traditional games and visual and performing arts. They also have the right to maintain, control, protect and develop their intellectual property over such cultural heritage, traditional knowledge and traditional cultural expressions. ”1 1 United Nations General Assembly, United Nations Declaration on the Rights of Indigenous Peoples (adopted 13 September 2007) A/RES/61/295 art 31. 1.
The UNDRIP not only recognises the rights of indigenous peoples over their expressions but also their rights to the intellectual property that governs them. 2 Traditional craftsmanship and culture is assumed to be part of the public domain, albeit when such work is misappropriated by a third party, it degrades the communal identity of the origin community. This is currently a widespread problem, an example being designer Jean-Paul Gautier using Maori patterns in his campaigns without crediting Maori as the origin source, thereby allowing him to unfairly benefit. 3 Mutilation and distortion of traditional expression is also a rampant problem. In 2013, Nike created women's sportswear that incorporated traditional polynesian tattoos, offending the Samoan community that reserved these patterns solely for the men of the community and Nike’s product disregarded customary rules.
4 The crux of the issue is to maintain a balance in protecting the rights of the communities and their expressions that prevent misuse , but to also ensure there is no atrophy culture. 5
2. Patterns, Protections and the Law
Under the Indian Copyright Act, 1957, Section 2(d) defines author as the person who creates the original work and is identifiable. 6 The Act further requires originality and a fixed material form to enable protection. TCEs struggle to fit this description as they are intergenerational, collectively authored with no singular author source or fixed documentation. The Indian Designs Act, 2000 governs the registration and protection of original designs, requiring the design (features of shape, configuration, pattern,ornamentation or composition of lines or colors applied to a product that appeal to the eye) to be original and registered to be granted protection.
7 These legal provisions are once again a vacuum where the rights of traditional artisanal communities disappear, as traditional designs fail to meet the requirement of 2Lily Martinet, ‘Traditional Cultural Expressions and International Intellectual Property Law’ (2019) 47(1) International Journal of Legal Information 6–12 accessed 2 November 2025 3 Mika Young, ‘Tā Moko and the Cultural Politics of Appropriation’ (2018) Sites: New Series 15(2) 1–15 accessed 5 November 2025 4 Aleni Sofara, ‘Traditional Knowledge in Samoa: At Risk of Being Lost’ in WIPO-WTO Colloquium Papers (2017) 8, 91–100 accessed 5 November 2025 5 Lily Martinet, ‘Traditional Cultural Expressions and International Intellectual Property Law’ (2019) 47(1) International Journal of Legal Information 6–12 accessed 2 November 2025
6 Indian Copyright Act 1957
7 Designs Act 2000
originality because they are historically transmitted and the lack of knowledge amongst rural artisans obstruct the ability to navigate the bureaucratic system of registration that requires an application, fee and technical awareness. While most Intellectual Property protections are individualized, with aspects like copyright also aiming to incentivise innovation, Geographical Indicators (GI) are based on collective rights. GI labels a product along with its characteristics as originating from a certain region, acknowledging collective and communal traditions and skills gathered through years. They allow for small-scale traditional artisans to enhance their reputations and sell directly to the consumer, also benefiting consumers by providing them with assurances of authenticity.
8 However, GI is only granted in cases where the TCEs is associated with one particular region, not if its sources are scattered. 9 That is a concern, as a community is not always one physical location, but a shared resource of custom and tradition. This could operate unfairly when a person from the physical community moves outside the geographical location. If a weaver of Kanjeevaram silk were to move locations, he may wish to pursue his art wherever he goes,but yet he is no longer a weaver of ‘Kanjeevaram’ silk and will not be protected. 10 Further, to gain protection under a GI, the TCE must also enjoy a reputation as GI signifies only the source of a good and cannot protect the information. For example, the traditional handloom of Kinnal (Karnataka) has been recently documented and the art is organized.
However, Kotpad Handloom Fabric has been traditionally established and a GI tag will not be effective for the former as it does not enjoy a long standing reputation like Kotpad Handloom Fabric does. Another example to be considered is Pochampally Ikat which involves a laborious process of weaving and human skill. Since the knowledge required for production is public, it may be manufactured anywhere, the uniqueness of the product and its quality depends solely on the skill of the weaver. Only in these cases will a GI tag be able to afford considerable protection as consumers may want to buy the products that are manufactured in that region, using a particular skill that produces distinct features related to the art of that area. While protecting the human owner, it is important to also protect the skill and knowledge.
The Ministry of Health and Family Welfare, along with the National Institution of Science 8 DR Downes, ‘How Intellectual Property Could be a Tool to Protect Traditional Knowledge’ (2000) 25 Columbia Journal of Environmental Law 253 9 Shivani Singhal, ‘Geographical Indications and Traditional Knowledge’ (2008) 3 Journal of Intellectual Property Law & Practice 732‐738 accessed 4 November 2025 10 M Liebl and T Roy, ‘Handmade in India: Traditional Craft Skills in a Changing World’ in JM Finger et al (eds), Poor People’s Knowledge (World Bank, Washington 2004) Communication and Indian Systems of Medicine and Homeopathy has developed a digital library of traditional knowledge11 mainly to prevent ‘poaching’ of firms that attempt to obtain parents on traditional Indian substances. Perhaps with time, a similar innovation can be afforded with relation to TCEs and crafts.
Until then, it's left to these communities to depend on secrecy as their very own form of IPR, risking the extinction of cultural expression. 12
3. Inspiration v Infringement
The concept of Idea versus Expression is central to Copyright as the Act protects expressions but not ideas. Inspiration refers to drawing from or being influenced by ideas without copying the expression. Infringement means to copy a substantial portion of someone else's protected expression and not a broad idea. The question this section attempts to tackle is when does use of TCEs knowledge that is in the public domain shift from mere inspiration to infringement in the context of AI generated designs and large fashion industries. Reiterating the fact that TCEs lack a single identifiable author and formal copyright, it brings to light the fragile demarcation of inspiration and infringement. Commodification by a third party in the form of derivative, adapted or inspired works is a cause of conflict.
13 To understand the argument, we must first clarify that every expression is influenced by an idea that is not unique to one. John Dewey defined art as an experience where creation and the appreciation of art is integrated with how we perceive ourselves and deal with the world around us. 14 Here, the introduction of AI complicates the distinction of inspiration/infringement further. AI produced its patterns/outputs from datasets that collect scraps of information over time, resulting in an output that may have minor alterations that appear to be ‘transformative’ but it strips away or misrepresents the cultural meaning that originated from the marginalized artisans.
These artisanal communities then face Epistemic Injustice, a term coined by Miranda Fricker, where their authority over 11 Traditional Knowledge Digital Library, Government of India accessed 2 November 2025 12 M Liebl and T Roy, ‘Handmade in India: Traditional Craft Skills in a Changing World’ in JM Finger et al. (eds), Poor People’s Knowledge (World Bank, Washington, 2004). 13 Gunjan Arora, ‘Preservation or Protection? The Intellectual Property Debate Surrounding Traditional Cultural Expressions’ (2025) Harvard International Law Journal accessed 4 November 2025
14 CHEN Jun and Danny Friedmann, ‘Protecting Sacred Art and Identity – From Intellectual Property to
Traditional Cultural Expressions’ (2025) Harvard International Law Journal accessed 7 November 2025 their cultural knowledge is ignored (testimonial injustice) and their interpretive frames are absent in the design outputs (hermeneutical injustice)15 If a pattern is modified superficially, but misrepresents the cultural meaning, it constitutes as a moral rights infringement. Since these AI design models are predominantly used by large industries, it enables unjust profit without any attribution or benefit sharing to the TCE communities, creating the distinction between ethical inspiration and exploitative infringement.
A counter under the idea of ‘Cultural Cosmopolitanism’ could be argued that not all appropriation is disrespectful, as the mixing of cultural elements is important for human flourishing and can be perceived as an opportunity for connection and empathy rather than allowing one group to restrict its culture within themselves. 16 In retaliation, we will use John Locke’s Theory of Property that states that a person owns their labour and adds that labour whenever they appropriate a thing from the Commons. If another takes the object the former person appropriated, he also takes the labour the first person added. This taking of labour is a harm. 17 A second argument is the Personality Theory linking the creator's expression to his personality.
18 In this scenario, AI when reproducing designs could risk undermining the authenticity of the original pattern, causing harm to the community's identity. 19
4. The Transformative Dilemma
The central dispute that arises when AI models produce modified outputs based on datasets of collected TCEs is whether the final output is transformative enough to avoid infringement. The case of Campbell v Acuff-Rose Music Inc. (1994) lays down the framework for ‘Transformative Use’ stating that when a work adds new purpose, meaning or message it becomes transformative.20 This understanding clashes with TCE contexts, as AI cannot create new cultural meaning as the meaning is socially embedded and inseparable from the community
15 Jackie Kay, Atoosa Kasirzadeh and Shakir Mohamed, ‘Epistemic Injustice in Generative AI’ (2024)
arXiv:2408.11441v1 accessed 4 November 2025 16 Kal Raustiala and Christopher Jon Sprigman, ‘Culture Appropriation and the Global Fashion Industry’ in David Tan, Jeanne Fromer and Dev Gangjee (eds), Fashion and Intellectual Property (Cambridge University Press,
2025) 316–341
17 Seana Valentine Shiffrin, ‘Lockean Arguments for Private Intellectual Property’ in Stephen R Munzer (ed), New Essays in the Legal and Political Theory of Property (Cambridge University Press, 2001) 138
18 Margaret J Radin, ‘Property and Personhood’ (1982) 34 Stanford Law Review 957.
19Kal Raustiala and Christopher Jon Sprigman, ‘Culture Appropriation and the Global Fashion Industry’ in David Tan, Jeanne Fromer and Dev Gangjee (eds), Fashion and Intellectual Property (Cambridge University Press,
2025) 316–341
20 Campbell v Acuff-Rose Music Inc, 510 US 569 [1994]
and its identity. Transformative use protects innovation, whereas AI merely produces a stylistic variation.
4.1 Colonial Codes in Algorithmic Design
It is pivotal to understand why allowing AI and Large fashion industries to interact with TCE is not a neutral act of inspiration but resonant with historical colonial practices of extraction, knowledge erasure and cultural domination. Spivak's definition of ‘Epistemic Violence’ through which dominant groups silence the subalterns can be applied to understand how AI designs erase the original communities history and impose a commercial narrative that allows the larger dominant fashion market to gain. 21 AI systems are designed to extract and manipulate information, this aesthetic sanitization removing historical, cultural and community specific identity embedded in the work.
The configuration of AI systems revolve around euro centric ideas of design and therefore the expressions and identities of marginalized communities are filtered through algorithms that fail to recognise their cultural frames of reference, further risking homogenisation of style. In view of Homi Bhabha's Of Mimicry and Man, AI mimics traditional designs and this mimicry (almost the same, but not quite) transforms into an uncertainty which establishes the ‘colonial’ subject as a “partial presence” meaning both incomplete and virtual. The success of the ‘colonial’ appropriation depends on the proliferation of inappropriate objects that ensure its strategic failure, causing this mimicry to be both a resemblance and menace. 22 This algorithmic monopolisation by powerful corporations conflicts with international norms of inclusive cultural participation.
23 Thus the argument holds that deriving traditional artisans of their collective rights causes material and cultural harm similar to colonial extraction and the asymmetrical gain of dominant actors.
4.2 Prada-Kolhapuri Controversy
A PIL was filed in 2025 in the Bombay High Court alleging that Prada replicated the aesthetic style of the Kolhapuri chappals at their Milan Fashion Week. The Kolhapuri chappals are GI protected and belong to the traditional artisans of the area. The petition filed for injunctive relief and compensation of the artisans but the Court dismissed it on the grounds that under GI the misuse of passing off of the GI name is actionable but not necessarily the aesthetic feature. 21 Gayatri Chakravorty Spivak, ‘Can the Subaltern Speak?’ in Cary Nelson and Lawrence Grossberg (eds), Marxism and the Interpretation of Culture (University of Illinois Press, 1988) 271–313
22 Homi K Bhabha, Of Mimicry and Man (Routledge 1994) 153–154
23 L M’Baye, AI, Fashion, and Cultural Appropriation (2024)
Prada’s avoidance of the label ‘Kolhapuri’ undercut any straight GI claim. The Amar Nath Sehgal v Union of India case emphasized on the moral rights protection encapsulated in Section
57 of the Indian Copyright Act and read it expansively to address the protection of cultural
heritage. 24 However, moral rights presume an identifiable author, making it difficult to protect moral rights when the artists are dispersed, as in the case of Kolhapuri artisans. The Prada- Kolhapuri controversy highlights the precedence of the statutory and procedural architecture of the law over economic and cultural harms, as IPR laws assume all rights holders to fit into a standard description. Relating this to Spivak and Fricker's epistemic harms, it's an obvious example of the marginalization of artisans by the dominant industry, where their crafts are rendered to mere aesthetics and stripped of their cultural significance. If humans are appropriating TCEs, it's indeed worrying what the use of AI will do, with less accountability and transparency and production on an industrial scale with no attributions to the original community.
AI could exacerbate the issue and turn episodic harm into systemic dispossession across various TCEs. 25 Since AI models and datasets operate cross-nations, it further complicates the jurisdictional aspects of available Intellectual Property protections. In an environment where fashion outpaces the work it borrows from, reinventing the law is not simply a policy issue, it's about protecting cultural identity. 26
5. Shadow Market Exploitation
This section aims to highlight that AI systems could create a parallel, unregulated market, that is exempted from legal scrutiny as it operates as a dataset and not a physical body, causing infringement that is far greater than economic harm. This market is an archive of scraps and no documentation of sources, with TCEs acting as mere inputs of the dataset. Once this archive is created, any industry or design would be able to access it without knowing the origin of the mutilated or collaborated TCEs. The current laws lack the legal provisions to address such a market, as they all deal with appropriation upon output, but here the appropriation happens
24 Amar Nath Sehgal v Union of India [2005] 30 PTC 253 (Del HC)
25 Marcelo Pasetti, James William Santos, Nicholas Kluge Corrêa & Nythamar de Oliveira, ‘Technical, legal, and ethical challenges of generative artificial intelligence: an analysis of the governance of training data and copyrights’ (2025) 5 Discover Artificial Intelligence 193 accessed
14 November 2025
26 Resham Jha, ‘Toe to Toe
Kolhapuri Chappals vs. Prada’s Luxury Strut’ Fashion Law Journal (11 November 2025) accessed November 2025 upstream, where TCEs are added to datasets that extract their aesthetics but leave behind their cultural meaning. TCEs act as a collateral without consent. In the case of Super Cassettes Industries v Myspace Inc and Anr, the Delhi High Court held that knowledge of infringement imposes a duty of removal, by this analogy AI datasets and their creators who knowingly collect TCEe and thereby authorise their downstream exploitation should be removed. 27 The collatorisation of TCEs though legally invisible, has consequences as it leads to an unjust increase in the value of the AI datasets without contributing to the original artisans who become hidden inputs. But the most significant harm that may arise from this remains substitution and not similarity.
As the AI extracts aesthetic style from these culturally significant expressions, it can go on to create ‘new’ patterns that occupy the same economic niche as the original tradition. The concept of Indigenous Data Sovereignty asserts the inherent rights of indigenous communities to govern the data about their territories, communities and resources. This concept finds its roots in the history of data extraction and exploitation, where information about indigenous people were collected without their consent and often misinterpreted or used to their detriment. These AI datasets could amplify these historical harms, making it a digital colonizer that perpetuates cultural erosion and deepening of social economic hierarchies. The same AI datasets could on the flipside become a platform for the flourishing of marginalized and indigenous TCEs.
But the core paradox remains that marginalized communities should engage with AI to ensure representation and avoid marginalization, yet risk losing sovereignty over their knowledge and data through this engagement. 28 The structural gap in the legal framework further exacerbates this dilemma.
6. Tradition in the International Lens
6.1 Australia
In Milpurrurru v Indofurn Pty Ltd the Federal Court of Australia held that aboriginal artists designs were being reproduced on carpets manufactured overseas without consent. The Court here awarded damages not just for economic loss, but also for cultural hurt, recognising the 27 Amrutha, ‘Case Comment: MySpace Inc v Super Cassettes Industries Ltd – Landmark Judgement Regarding Safe Harbour Immunity of Intermediaries and Interpretation of Various Provisions of IT Act 2000 and Copyright Act 1957’ (2022) 3(3) Journal of Legal Research & Juridical Sciences accessed 10 November 2025
28 Sustainability Directory, ‘Ethical Frameworks for AI and Indigenous Knowledge’ (Scenario)
accessed 11 November 2025 communal harm that affected the artists, offending their relationship and identity with their art.
29 The judgement acknowledged that the designs had elements that were sacred to the
community and reproduction of the same incorporating changes harmed the cultural significance and caused humiliation. Though normative copyright laws recognize individual authorship, the Court here acknowledged communal custodianship and evaluated the case on communal and traditional harm. However, though this case accounted for customs, the law does not formally recognise continued custodianship over future productions. This case illustrates a structural gap, where while the courts can recognise cultural hurt there are no institutional mechanisms that allow regulation.
6.2 The European Union
The EU design requires novelty and individual character, making it hard for communal or shared TCEs to fit this requirement. It's modelled on Western notions of individual creativity and rarely acknowledged communal creativity, thereby providing little to no recourse for traditional artisanal communities unless they have been formally recognised. The observation derived is that even developed IP systems prioritize individual authorship propagating a global structural bias against TCEs.
7 Weaving a Legal Framework
The Sui Generis Model Protection of TCEs is a global problem, but differs within each country’s environment. Global frameworks do not consider or provide complete remedies for the protection of TCEs that fulfill the standards that resonate with different communities and their needs. This makes the introduction of domestic Sui Generis models crucial, aligning domestic laws within the international framework. A system that enables and recognizes collective rights and authorship is necessary to protect the sovereignty of traditional artisanal communities. There has to be equitable benefit sharing schemes so that communities are not unfairly exploited of their knowledge and skill. There is a need to strengthen cultural legitimacy by incorporating customary laws via participation of these communities in the drafting of the system.30
29 Milpurrurru v Indofurn Pty Ltd [1994] 49 FCR 191
30 Aiswariya Venugopal B, ‘Traditional Knowledge and Its Protection Under Indian Intellectual Property Laws: Challenges and Prospects’ (2025) Journal of Law & Legal Research Development 2(3) 1–4 accessed 13 November 2025 In the context of AI, there should be a mandatory disclosure of sources used in datasets and model training and in the case of generated designs there should be watermarks embedded that inform the consumer about the origin source, and if the product is generated using AI to ensure full transparency and enable accountability. The structure of GI should ideally be extended to Copyright to recognise communal and dispersed authorship allowing collective rights and protections. This follows the reasoning in Milpurrurru v Indofurn Pty Ltd where courts recognised communal cultural harm.
Such a hybrid model will help avoid the current discrepancies in the legal framework and empower marginalized artisanal communities and prevent their exploitation by global industries and AI, reinforcing their authority over their communal knowledge and skill. These measures will align with UNDRIP and UNESCO conventions that impose duties on states to protect their cultural heritage and the related communities rights. Further, most Intellectual Property laws require the author or the person infringing to be human, allowing AI generated work to fall into a grey area. Leaving the use of AI unchecked could result in a huge increase in copyright issues. Though AI works faster than human labour, most of its creations are based on pre-existing patterns and knowledge, hence boosting the counterfeit market.
31 The legal framework could also choose to reinvent AI systems by protecting creativity instead of imitation, this might be hard to achieve as AI systems have no consciousness and operate merely on datasets that reproduce inputs in various forms and cannot therefore substantially contribute to innovation. But being able to incorporate AI and making sure AI systems and human moral rights can coexist is imperative in a fast paced world that is constantly evolving. 32 The base line of the framework should be to protect culture and heritage, promote collective rights and ensure regulation and oversight.
8. The Last Stitch
Taken together, this research supports a broader claim. It calls for a conceptual shift that resolves the structural conflict between two epistemic systems. One that is defined by cultural 31 Christophe Geiger and Vincenzo Iaia, ‘Fashion, Intellectual Property and Freedom of Artistic Expression in the Age of Metaverse and AI: A Digital Constitutionalist‐Approach’ (2024) European Intellectual Property Review 46(9) 555‐570 accessed 15 November 2025 32 Marie Malaurie‐Vignal, ‘Could Fashion Copies Become Lawful? ’ (2018) 13 Journal of Intellectual Property Law & Practice 657–663 accessed 13 November 2025 stewardship and common authorship and the other is a global industrial economy built on data extraction and distribution. Intellectual property laws are not fully equipped yet to mediate the boundaries between the two, with gaps in its understanding of authorship and infringement.
A way forward requires the law to recognise and categorise collective ownership and rights, account for not just economic harm but also cultural and epistemic harm that occurs from data extraction by AIs and larger fashion houses and giving communities power and agency over their TCEs, knowledge and skill and to enable global benefit sharing. As AI continues to gain dominance, the question is not if and whether the legal systems should intervene rather it is how they should intervene. Will Intellectual Property regimes continue to allow algorithmic exploitation at the expense of cultural integrity or can it embrace cultural and traditional identities and TCEs by awarding them recognition and rights?