Courtesy of Alena Stepanova

We are living at the height of what some might consider “The Fourth Industrial Revolution”, a period marked by an increased reliance on automation, data exchange and connectivity. Whether we like it or not, AI will become a significant part of our lives and culture (if it hasn’t already unbeknownst to you), where non-physical tasks can be performed significantly faster, easier, and more efficiently. Therefore, it’s a no-brainer that the fashion industry is now employing this tool to enhance accuracy and effectiveness in different stages of the production process, such as marketing, design, and supply chain management. In this feature, we will expand this discussion by first examining how AI is applied to the fashion industry in conjunction with other modern technologies, and secondly, what AI means for the future of fashion sustainability.

Courtesy of Matilde Mariano

How is AI transforming the fashion industry?

AI has opened a wide array of performance-based possibilities and has been transforming the ways industries operate. In the fashion industry, AI can be applied to every stage, from design to production to marketing and sales and streamline each of these processes. While the more obvious use of AI in the fashion business is to generate effective marketing strategies and advertising campaigns to target customers, in the design stage, AI algorithms can provide a general creative blueprint for designers, based on predicted trends and popularity. In an ironic way, fashion designers can now use AI to get “closer” to their consumers without physical interactions. For example, AI-embedded wearable products are now available that capture body measurement and track user motion. These data can be stored and analyzed to decide on future trend specifics such as color, fabric, and silhouette to create personalized clothing. This type of bespoke service that pre-confirms the customer’s preferences and tastes can prevent wasting excess material while at the same time boosting sales.

Perhaps a larger scale of waste reduction is possible at the supply chain stage. AI models containing data of historical inventory levels/sales, market trends, social media posts and environmental impacts that can be used to inform fashion brands to make optimized decisions for their future products. For example, AI can be used to predict the precise quantity of pieces to produce to satisfy market needs while preventing deadstock. Thus, AI could feasibly be used to provide solutions as to where and how to use resources more ethically and streamlining the entire production process to mitigate energy consumption.

Though AI can provide solutions to challenges faced by the fashion industry such as overproduction, there are potential cons revolving around diminishing creativity and individual expression, and the promotion of faster production and consumerism. The former can occur when brands collectively rely on algorithms to determine designs that fit the market. The result of this is the lack of innovation and more homogenized fashion collections. The latter concerns using AI to target audiences for the sake of creating high demands to drive sales leading to overconsumption. Perhaps the greatest caveat of all however, is the shocking amount of energy is required to train and run AI models in general, which in turn offsets the benefit of AI to solve sustainable fashion problems.

Courtesy of Aetheria Designs

Coupling AI with existing modern technologies to further streamline processes

Technologies are often interdependent and need to be compatible and function in concert with each other to be useful. This holds true for AI as it is integrated or paired with existing technologies to enhance processing time and accuracy. For example, AI relies on cloud computing where vast number of on-demand resources such as data storage and analytics are readily available. In the fashion design and production stages, cloud-based AI systems can assist designers by generating 3D mock ups of clothing to give them a better idea what the pieces will look like. The approach can also be used to analyze consumer data to optimize the scale of the production based on the fluctuating market demand, which overall saves businesses money by reducing upfront investment and the likelihood of overproduction.

Another example is the marriage of AI and the Internet of Things (IoT). IoT is essentially a network where physical objects in any type of system can “talk” to each other and the environment around them by exchanging data through sensors that are connected to the internet. These data are highly valuable in the fashion industry because it provides visibility in the supply chain process. When it comes to sustainable fashion, this is beneficial because it could streamline production. Data gathered from IoT can also train AI to determine when there is an excess of inventory or when the stock is low, allowing fashion brands to make accurate judgement of production quantity in real-time. Collectively, these synergistic interactions can ease production time and resources and hence, reducing material waste.

A more customer-oriented use of AI is to combine it with augmented reality (AR) technology. An sample of this is Farfetch’s proposed “Store of the Future”, where Smart Mirrors are installed for customers to virtually browse and try on different clothing, and can share their entire shopping experience digitally with people outside the store. The AR technology will give customers a more robust interactive shopping experience offline, where customers might feel less pressure to make purchasing decisions right away. The Smart Mirror can gather these digital footprints, analyzed by AI-powered algorithms to help optimize and improve customer services.

Courtesy of Laura Valentino

The future of AI-driven sustainable fashion will depend on developers and users

As we collect more and more data and AI technology becomes more refined, it is possible that the power of this tool could provide tangible solutions for attaining fashion sustainability. Brands who built their business model with a sustainable mindset, and value environmental consciousness will likely find AI particularly beneficial. However, the harsh reality is we are living in a market economy that is highly profit driven. Business models like ones in Fast Fashion will likely rather encourage overconsumption to break sales records. Therefore, AI will only be as benevolent as intended by the user.

A bigger problem with AI models, which NJAL Features has previously touched upon, is that we have yet to figure out how to mitigate the amount energy (i.e. electricity) required to train and process the technology. The demand for AI usage skyrocketed in recent years – by 2027, an estimate of between 80 to 130 terawatt-hours of electricity is predicted to be consumed annually for AI maintenance, which is enough to power a small European country like Netherlands. Further, the amount of CO2 emission is an even a greater potential concern. Recent reports have estimated AI systems such as ChatGPT emit about 3.82 metric tons CO 2 per day, which is equivalent to someone driving 10,000 miles in a gasoline-powered vehicle. Since there is not yet a transparency requirement or reporting system like that of transportation (e.g., air travel, trains, automobiles etc.), AI companies and developers can report whatever or however they want regarding energy consumption and waste production.

Courtesy of Metamorphix

As mentioned, AI technology will become a norm of the future, just like the internet and mobile phones. Data scientists and Institutions such as the Lincoln Laboratory Supercomputing Center are currently developing methods to reduce the amount of energy required to build and process AI models. These methods include capping energy usage in general, adjusting data center operations and further optimizing AI accuracy by building upon previous models. Early studies showed that the latter method could reduce as much as 80% of energy consumption in training AI. These efforts will likely have a knock-on affect for how AI is used in the fashion industry.

However, now, the answers to whether AI can truly benefit sustainable fashion or set us back even further remains unknown. As this technology is still considered at its infancy, there is a lot needed on the research front to determine whether it can be of help or of hindrance with regards to the energy crises we are facing. At least some Researchers have found that AI usage emits less carbon footprint than humans in some activities, and this perhaps will become more realistic in the future when the tedious AI training process is fulfilled. Furthermore, like many things, a cost to benefit assessment and the cooperation of the entire industry is crucial; AI can be used for good or bad, but it is in our hands to determine so.

Visit NJAL I Designers to discover innovative designers working at the intersection of craftstmanship and new technologies.