The use of Artificial Intelligence (hereinafter, “AI”) has burst into the market, gradually transforming our lives and revolutionizing the approach to Intellectual Property (hereinafter, “IP”) in view of the legal challenges put forward by the strong impact and development of new technologies.
In the fashion industry, the use of AI is progressively becoming a fundamental factor for companies to remain competitive, and its growing application raises new questions from an IP perspective, among others. Increasingly, brands are breaking with tradition and going digital, applying AI algorithms and tools not only for the creation of their models but also for the management of numerous industrial processes.
Particularly in 2023, there was an increase in attention — if not concern — regarding the need to find answers to the new legal issues posed by the use of artificial intelligence, including generative AI (hereinafter, “Generative AI”), especially in the United States of America (hereinafter, “USA”) and the European Union (hereinafter “EU”), where the use of these tools continued to grow exponentially.
At the beginning of that year, the influential Harvard Business Review [1] questioned what lies behind the major new advantages offered by Generative AI in terms of content and creations. It argued that the possibility of generating multiple high-quality designs, as if by "magic" and at a staggering speed, were not truly "magical" results, but purely derived from pre-existing material with serious implications for IP rights in creative industries such as art, design, publishing, advertising, and, of course, fashion.
In the USA we see interesting progress in the IP sector by lawyers as well as by the registration offices and courts, all of whom need to learn how to manage new cases and establish guidelines for resolving legal issues arising from the use of Generative AI by registration applicants, particularly regarding copyright. It is not surprising to see this pioneering and practical effort in the USA, a jurisdiction not only at the forefront of innovation but also home to the tenacious Common Law legal tradition, whose school of thought seeks to address legal problems rather than merely elaborating a defined body of determined rules.
Many lawyers find themselves dealing for the first time with requests from clients seeking to register copyrights for artistic creations, works and designs made using Generative AI. The difficulties in this regard were many: how to explain to a company or a designer that it might not be possible to protect their creation made with Generative AI? To what extent can such a creation be protected by copyright? And, if applicable, who should be considered the author(s) of such works?
On March 16, 2023, the U.S. Copyright Office (USCO) issued the first Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence, which addresses, among other issues: (i) whether a work produced by generative AI is protected by copyright; (ii) whether a work consisting of both human-created and AI-generated material can be registered as a work protected by copyright; and (iii) what information applicants wishing to register these types of works must provide to the USCO to seek copyright protection.
It was through a significant process of jurisprudential reflection that these initial questions began to be answered, culminating in several clarifying rulings (such as Thaler v. Perlmutter [2] and Kristina Kashtanova [3]) that provide the basic legal criteria for determining protection and identifying copyright ownership in works created using Generative AI. In summary, these criteria are twofold: (i) a human expression and (ii) a final creative control of the author. Only if the new creation reflects these aspects can it be considered a work protected by copyright. Furthermore, the USCO imposes an obligation on applicants to disclose whether their works include any content produced through Generative AI.
At the same time, the EU is also making increasing progress in the development and use of AI, logically dealing with the same IP issues. However, the approach to these problems is handled under a Civil Law system that seeks adequate legislation — a challenge that seems perhaps even greater given the staggering speed at which technology evolves.
Both the EU regimes and international copyright protection systems align with the U.S. position regarding the principle that originality must stem from the human mind, thereby rejecting machine authorship. In other countries, such as Canada and Australia, an "intellectual effort" and "original expression" by a physical person are also required for rights protection.
In a peculiar move, the United Kingdom (hereinafter “UK”) adopts an atypical position whereby, under its Copyright, Designs and Patents Act, a "computer-generated" work may be protected if there is a human being who has made the arrangements necessary for its creation. This means that if an AI tool has created a design, a court could consider the human who configured or executed the program as the creator.
The UK’s stance, in particular, raises thorny questions about ownership — that is, who should be recognized as the owner of a work, a design or any other form of artistic creation.
The question of ownership is of immense relevance in the fashion sector, considering that human designers have always been the artistic inspiration behind collections, yet they are increasingly using the help of Generative AI to create their designs and models. In this sense, even if we accept that AI-assisted fashion is eligible for copyright, the legal dilemma of authorship remains: who is the author? Is it the person who entered the prompt? The developer who trained the AI model? The brand that financed it? Or the company that owns the AI model?
As illustrated by examples provided by Your Fashion Law Guide[4], imagine a luxury fashion brand that uses an AI tool to generate complete dress silhouettes and then simply selects the final designs. In the US, this might not be sufficient to recognize authorship over said designs. If there is no human intervention involving substantial editing or arrangement of the designs by the designer, the collection could fall outside the scope of copyright protection, resulting in it entering the public domain. Conversely, in the UK, a design originated entirely by AI could be protected by naming the AI operator as the author. Imagine a British designer who writes the code and configuration for an AI-based fashion application; under British law, this could make her the author.
According to Mrudula Kale, fashion law now stands at a crossroads. On one hand, we have the rigid formalism of “human authorship or nothing”; on the other, there is a vague technological optimism that trusts the courts will "figure it out." However, between these two extremes lies a more thoughtful future — one in which laws evolve to recognize the role of AI without erasing human creativity, and where the rights of creators, designers, and workers are protected even in a machine-assisted world.
In the current fashion landscape, we believe that human creativity remains the central pillar of IP in the face of the rise of generative AI. Although technology optimizes design and production, emerging regulations must seek a balance: integrating algorithmic efficiency without stripping protection from the talent of creators.
In conclusion, the regulatory framework must incentivize innovation while preventing automation from facilitating the exploitation of third-party works. Copyright protection should prioritize those who, while using AI as a tool, maintain the human essence as the driving force behind success and brand value.
Article originally published on Mundo ASIPI Magazine
