(H)AI FASHION REVOLUTION? In search of a solution to the problem of fast fashion.
The path from personalized tailoring through the “copy-paste” fashion industry has definitely been one of the most winding roads ever seen in the history of mankind. As the world started to open up, and shrink due to both wider travel opportunities, and the Internet occurrence, the clothing production had to accelerate its pace as well. In consequence, the fashion industry has outperformed almost all the other sectors in the consumption race, producing at the same time about 10% of humanity’s carbon emissions[1]. That places the textile sector at the inglorious second place of the largest use of the world’s water supply, polluting 20% of all industrial water worldwide. According to the World Bank report, if we continue generating global consumption at the present level, that might result in the total amount of 11 million tonnes of junk per day by 2100[2].
Needless to say, the issues caused by the modern clothing sector have spilled over not only the environment, but also on society itself. The problem of fast fashion has become more popular within the past decade. It refers to the similar character of fast food production, that is based on mass-production of junky goods at the most possible low cost, with selling them by relatively low price as well. As their main goal is to follow recent trends from the catwalks, and transfer them to their own design, the secret of fast-fashion brands popularity lies in two, general conditions — the opportunity to taste a little luxury for little money. The beginning of the phenomenon of fast fashion emerged in the late 1990s, when shopping has started to be considered as a form of entertainment. According to Rudradeb Mitra, thanks to economies of scale, average people could afford much more things than before. As a result,
their massive consumption has formed the main shaft of modern economies. Additionally, research conducted by IBM analysed the general amount of thrown away clothes. They found that of 2,000 consumers surveyed by OnePoll, almost 20% lands in the landfill [3]. Years of technological development has powered the above said demand for the biggest possible possession [4]
Notwithstanding, it is also technology that might take control of the situation. Thanks to the fast development of AI driven mechanisms, many solutions to tackle the problem has showed up. Opposite to human tendency to depict artificial intelligence as a serious threat to mankind, on a scale that has never been seen before it is technology that can be our only remedy for the fast fashion global destruction. AI forced transformation has already started to impact the fashion industry as a whole — from the very first design, through all of the actions undertaken within the manufacturing process, towards getting through to the customer, by selling and proper marketing. Nowadays, AI can tackle the tasks that humans have been doing for many years. Using technologies based on Machine and Deep Learning, Visual Recognition and Natural Language Processing, AI can process larger quantities of data than a man could do, extracting from it exactly what is needed to answer a certain question or make a proper choice. This technology can help both brand owners as well as designers to combine two most precious values — achieving
sustainability production in a short time. From market prediction and demand preparation, AI is making all the chains more sustainable. It can also be found in the McKinsey report that AI can reduce errors by 50% and overall stock amounts by 20–50% [5].
First, it should be underlined that the most significant shift done by AI solutions is associated with personalization. The only thing fast fashion can do is to dress people the same, denying what is most important in the clothing industry — individual, original self-expression. There are already some companies that use AI to enhance brand-client relationships. For example, Heuritech’s invention is based on scanning around 3 million Instagram photos daily from different users to collect data and fit a customer to the specific brand [6]. As a result, it gives back the investment for the brand, minimizing the overstock, optimizing turnover and improving production at the same time. The Project CeCe works quite similarly to the above mentioned Heuritech solution. Established in Netherlands, online marketplace for sustainable fashion allows its customers to upload photos of dream outfits, and search for similar, but less harmful for environment, fashion designs [7].
Although meeting personal demands may provide a better relation with customers, their needs are very often directly bonded with what’s currently en vogue. Nowadays, fast fashion brands like Zara or H&M instead of having 2 or 3 collections per year, have at least 2 collections for one season, which causes more harm than good. Thus, it is important to be aware of trends or, what’s even better, to predict them in advance. In that matter, AI is a perfect tool to do so. If used sensibly, it can not only ensure profitable outcomes for the company, but also lead to reducing waste in our environment. As a perfect example should serve Sorabel company, that uses AI to foster an inventory risk at the low level while supplying items at relatively cheap prices. Their AI forced engine forecasts trends, which might probably be best to sell out [8]. This type of using AI innovations aaplies not only to small retailers, but also to some industry giants like Walmart or Amazon. Likewise Sorabel, those two companies use machine learning to identify most probable customers’ choices to be made in the short future [9].
Despite the strongly contaminating process of producing a specific cloth, it is also the delivery that causes a big, environmental harm. Most of the factories are based in the eastern part of the globe, while the biggest demand for cheap fashion might be met in the West. According to the social media impact on raising people’s awareness of the idea of sustainable development, for some of the customers it is really important to take over the control of their buying choices. Thus, software startups like Trustrace help to visualize supply chains that are mapped with company-specific sustainability metrics and powered with supplier data. It consists of 5 elements that altogether serve to trace the cloth way to the customer. One of the parts called T-CAT is even able to measure the product level environmental footprint of the product and after simplifying collected data, communicate the level of the footprint to the customer within different channels [10].
All of the above mentioned ideas can apply to the processes that include creating clothes from scratch. Recent study has shown however, that for the real implementation of sustainable development policy, this type of production, no matter how ecological, leads nowhere but to the beginning of the issue. Thus it seems that the most conveniable solution should be changing our modern economy from a linear one to a more circular type. Generally, it is based on three main principles: designing out the waste by choosing the natural resources instead of artificial ones, emphasizing the quality, durability and repairability of the item, and minimizing the effort to recycle the product. With leaving the linear system for good, we would have to abandon our current need to compete with each other — in the circular economy model there is no place for an exclusive mindset, and only those who can contribute to the system, will be able to take full advantage of it. Some elements of the circular economy philosophy have already been implemented into the biggest companies’ ecology-like ventures. However, the actions undertaken to run towards the sustainability requirements couldn’t be met without AI mechanisms — according to the United Nations Big Data for Sustainable Development report, new sources of data, new technologies, and new analytical approaches, if applied responsively, can enable more agile, efficient and evidence-based decision-making and can better measure progress on the Sustainable Development Goals (SDGs) in a way that is both inclusive and fair [11]. One of the most important ideas in that field has been implemented by thredUP company, that according to Forbes, is poised to capitalize on the growing $24 bilion second hand market through its
use of AI to bring efficiencies and scale to every area of its operations, while fueling the circular fashion trend among traditional retail brands with the launch of its “resale as a service” offering [12]. In other words, the company’s goal is to make second-hand clothes fashionable again. AI is being used here for many different tasks, such as pricing, personalization and styling services. Being more precise, the company uses image recognition to visual tagging and assigning attributions based on various cloth features.
Becoming sustainable is a really tough nut to crack. Although AI can support many bright ideas to cure environmental harm, there is still a lot to be done in people’s approach to fashion and buying itself. According to the PWC Global Artificial Intelligence research, we could eventually move to fully interactive and customized design and supply in which AI created mock-ups of garments are sold online, made in small batches using automated production, and subsequent changes are made to design based on user feedback [13]. Notwithstanding, without making customers aware of their choices, the shift won’t be done.
Links:
2 https://www.worldbank.org/en/news/feature/2013/10/30/global-waste-on-pace-to-triple
5 https://www.mckinsey.com/industries/retail/our-insights/renewed-optimism-for-the-fashion-industry#
6 https://www.heuritech.com/about-us/tech/
https://www.projectcece.co.uk/index/
10 https://trustrace.com/solution/
11 https://www.un.org/en/sections/issues-depth/big-data-sustainable-development/index.html
13 https://www.heuritech.com/fashion-brand-trend-forecasting/