The rise of A.I in fashion
Reading time: 05:00 minutes
In this article, we discuss how fashion is already adapting and embracing AI technology, and how this could impact the future of our creative business.
Technological innovations are already being implemented by fashion businesses throughout the supply chain to speed up manufacturing processes and to provide new improved levels of customer service. What will the rise of the machines mean to the creative roles in the Fashion industry?
Amazon is pioneering the development of AI in their business, with several startup projects that support both their ‘back of house’ and also customer facing angles of their empire. These vary from monitoring social media posts to determine the latest trend or popular item to its stylist app ‘Echo Look‘ that gives you feedback on your chosen outfit.
Pushing things further, an Amazon-backed team in San Francisco are working towards creating an AI fashion designer. The team at Lab 126 have created a program who’s algorithm can read a particular set of fashion images, for example, a trench coat, and then generate new items in a similar style from scratch. This would not only allow for manufacturing times of fast fashion copies to be drastically reduced, but it could also lead to the beginning of the end of a commercial designer role. An AI computer will generate far more ‘versions of’ or iterations, much faster than a human can. The possibility of reducing the speed to market will be a huge draw for many retailers, especially those whose businesses are driven by online sales and the consistent refreshing of the ‘new in’ page.
Nike has used data in a more complex way to create a range of clothing. They have fed the performance data of athletes into a knitting machine to develop garments adapted to the specific performance aspects, such as extra movement or ventilation. The data mapping is combined with their fly-knit technology to create high-performance products. This nine-piece capsule collection of menswear pieces was released last September and is part of a series of innovative data-driven projects Nike are working on.
In another part of the industry, trend-forecasting companies are also using data to drive intelligence, and are using this in incredibly innovative ways. By scanning inventory on online stores, the trend service Edited can see when an item is selling out, or if a particular colour is being brought across multiple retailers, determining real-time market trends that can then use as guides to chase into the competitive market for missed opportunities. Edited have pioneered this model and it has proven a vital tool for major retailers in a challenging climate.
AI is also being trialled and tested in fashion marketing and publications, to varying degrees of success. Stylist magazine ran a series of pieces in February this year created by an AI program called Articulo. By inputting keywords about a small brief into the computer, the machine-generated an article by scouring the Internet for relevant information and trends to formulate a view. It then created a 500-word piece on the set subject. Although generally the pieces were a little formulaic and lacking emotion, with a few grammatical errors, the structure and content were there. In an industry under cost and time pressures, will this AI technology replace part of the editorial landscape in the coming years? Let’s watch this space!
Social media has become arguably the biggest focus of fashion marketing, increasingly becoming more important than a traditional ad campaign. Instagram is often the most utilized, where influencers drive us to our next purchase as we aspire to the lifestyles their feeds create. The more pessimistic of us will already believe this is all an illusion, created for us to part with our hard earned cash; the perfect latte in the image is cold, and that tousled windswept hair took five hours of preening and teasing by a professional hairdresser. But what if it was a true simulation of a lifestyle we were buying into, generated by a computer?
@lilmiquela is listed on Instagram as a music artist, blogger and model living In LA, with Spanish and Brazilian heritage and has 886K flowers to date. She’s young and stylish, wears a variety of labels from Chanel to Off White, has a portfolio of magazine covers and an enviable lifestyle. In reality, she’s a computer-generated virtual avatar who’s creators are yet unknown. She’s amassed a huge following with her devotees affectionately calling themselves “miquelites”. Her avatar age is 19, so should we be worried that young people are buying into a simulated lifestyle that is quite literally ‘virtual reality’ and unobtainable? Or is the adaptation of technical advances into social media inevitable, and the boldest way to create a multi-talented product that can evolve in a time of rapidly changing tastes?
We are already experiencing the adaptation of AI technology into our everyday lives; from Siri to Alexa our homes can be managed via these intelligent devices. These can provide ease and convenience, making monotonous tasks simpler and taking away some of the daily chores, ordering milk when we run out, or dimming lights to improve energy efficiency. But algorithms drive AI devices, and algorithms work on patterns. The computer determines a pattern of behaviour and bases its decision from that pattern. By looking for repetition the program will create a formulaic response every time as machines cannot determine or feel emotion. This, therefore, excludes the possibility of greater choice; it will simply offer a more streamlined uniformity within the products created or offered. We will end up with the same milk in the fridge every week, the same shoes on our feet and the same clothes autonomously ordered. The freedom in fashion to express ourselves and make uninformed and emotional choices will be gone.
When adapted into the creative realms of the fashion industry, the choice could become less when the thought process and human emotion are removed. The choices made by the computer will be more limited, and therefore so will the products themselves. Or will these nuances eventually be built into computer learning and bring about a more sophisticated ‘illusion of choice’?