By Nathan Reiff
9 min read
Both the fashion industry and the emerging artificial intelligence (AI) space face a host of ethical dilemmas and considerations. In fashion, these concerns include ones related to labor practices, environmental impact and sustainability, representation, stereotype, and more. As a younger industry, AI’s ethical concerns are only more recently coming into focus, but they are similarly significant and include inherent bias in AI systems, transparency, replacement of human jobs, and intellectual property, among others. It is only natural, then, that the confluence of AI and fashion would similarly bring about a number of prominent ethical issues.
There is no doubt that the fashion industry has already been—and will in all likelihood continue to be—impacted by AI in profound ways. From reorganizing and streamlining supply chains to changing the way that designers conceptualize clothing to providing customers with new shopping and marketing experiences, the potential applications of AI tools in fashion are many. This technology has the capacity to reinvent fashion in many ways, including in some that are likely to increase transparency, to reap environmental and sustainability benefits, and to reduce the incidence of unfair labor practices.
Still, especially those involved in the merging of these two spaces but even everyday consumers as well should be aware of the ethical concerns at play when fashion and AI come together. Below, we explore some of the most salient ethical issues to be aware of in fashion and AI.
Privacy is a major concern for AI systems of all types. Most AI is able to accomplish its goals by exploring and analyzing a huge amount of data. When these data troves include personal information such as photographs, body data, purchasing habits and histories, personal interests, and more, individuals may unwittingly find themselves contributing to the growth of AI. When it comes to generative AI, this information may even make its way into new content that is generated. The field of AI is adapting to consumer concerns, with some of the most prominent platforms and developers experimenting with new security and privacy measures.
AI’s use of consumer data in the fashion industry can help companies to better anticipate aesthetic trends, to more accurately order products and stock shelves, to tailor marketing campaigns based on individual demographics, and much more. There is no doubt that AI can be useful in helping fashion companies to sell more products with less time and effort, and in helping customers more easily find the clothing they want. But there are important considerations about what kinds of data AI uses in order to accomplish these goals and how that data is employed or made available to the broader world.
Like the privacy issue, intellectual property is a significant issue for AI systems of all kinds, and particularly generative AI. Societies and legal systems are currently grappling with difficult questions in this area, including the dilemma of who owns the work created by generative AI. An argument could be made that this content is the property of the developer of the AI system. If it is created using data from other sources, those sources may also have an ownership claim to the content. Or it is also possible that no one person owns this new creation.
Generative AI is a powerful tool for fashion designers. Systems like Cala and Designovel are AI tools that help designers to create new items of clothing. AI makes it possible for designers to create life-like mockups of potential new items without having to resort to the costly use of materials, to putting in the labor of making a real-life prototype, and so on. In some cases, AI tools work in partnership with designers to provide new ideas or to realize a human artist’s vision. In others, AI can use prompts provided by a human to design a new item or even an entirely new line. The questions of intellectual property and ownership of these creations are especially urgent in these cases.
This is likely to play out in legal battles surrounding copyright in the AI/fashion landscape in the years to come. AI can play a part in creating new patterns, colors, and even materials, and it will be important to determine who has a claim to copyright ownership of these products and concepts. There is also currently a lack of a clear legal framework for ownership in these cases, which means that the stage is set for participants in the fashion industry to lose out on the financial benefits of copyright via theft, copying, and related practices.
AI is notoriously biased. Biases emerge based on a number of factors including the bias of a data set used to train an AI system and biases built into the system by developers. Unfortunately, bias in AI can lead to detrimental real-world effects such as discrimination in hiring practices.
When it comes to the confluence of fashion and AI, bias and discrimination are a significant risk. Studies have found that the fashion industry broadly holds biases in favor of white men and against minorities and women. There is also a long history of bias and discrimination based on shifting beauty standards and body types. It is no surprise, then, that when AI tools trained on data including these biases are asked to produce new content, this content often reflects those biases. Without properly intervening at the development or the execution levels, fashion companies utilizing AI tools may perpetuate these biases without being aware of them.
Throughout its history, the fashion world has increasingly relied upon new technology to help to speed up design and manufacturing processes. This has been possible thanks to inventions like the sewing machine and, more recently, various printing, cutting, and fusing machines, to use a few examples. In most cases, the incorporation of new tech into the manufacturing process has also resulted in the automation of some aspects of the process, thereby reducing the human workload, cutting costs, and ideally becoming more efficient.
One of the biggest concerns that people have about AI is whether it is likely to take over jobs occupied by humans. While the answer does seem to be yes, it’s not entirely clear to what extent this may be the case. The concern is very much a factor in the fashion world as well. If AI is developed that can effectively eliminate or dramatically reduce the need for people to be involved in the manufacturing process, what are the ethical concerns for designers and fashion companies involved in deciding whether or not to implement that tool? And should those companies have an obligation to consider the welfare of their employees in the process?
As we’ve seen, the fashion industry has an outsized and notorious history of pollution, waste, and environmental destruction. At the same time, AI also has its own related set of environmental concerns. AI tools require significant computing power and, therefore, energy. The computing power necessary for key AI landmarks doubled about every 3-4 months from 2012 to 2018. As this power need continues to increase, and as AI meets up with the fashion world in new ways, participants in the fashion space will be confronted with important decisions about how their use of AI impacts an already-challenging environmental situation. It may be possible that AI can help fashion to adopt more sustainable and environmentally-friendly practices. But that will likely only happen with dedicated work toward those goals.
These are just a few of the emerging ethical issues surrounding the intersection of AI and fashion. It’s likely that others will emerge as these two worlds continue to interact more and more. The legal gray area surrounding most things AI-related and a history of ethical concerns in the fashion world could mean that ethical issues are amplified in the years to come. On the other hand, the power of AI may help concerned participants inside and outside of the fashion industry to chart a new path forward that reduces some of these challenges as well.
Decrypt-a-cookie
This website or its third-party tools use cookies. Cookie policy By clicking the accept button, you agree to the use of cookies.