Creativity and Data: The Future of Fashion Design by Rafaella de Freitas

In August 2017 Amazon released a report[1] detailing how a team of programmers developed an Artificial Intelligence (AI) system that was able to generate new images of products from the input of product category (male or female tops, bottoms or shoes) and user style, determined by analysis or a number of images associated to the user such as online purchase history. One year later, and the potential of designer algorithms remains a trending topic.

The ability of an algorithm to generate designs is fascinating, but it also brings to attention questions of the value and purpose of fashion creation. Understanding whether there is intrinsic value in the creative process, or only in the final outcome? If the fashion industry is fully consumer-centric, or if is there space for the expression and creativity of designers? And determining what is valuable in a product: the sensory experience or only the image?

These are some of the questions that emerge with the introduction of a new method of design creation. The emergence of creative algorithms in fashion also brings to light the links between fashion and society: in the role of fashion in capturing and representing human experience, as well as its importance in how people understand themselves and others.

DESIGNING FROM DATA

The two main ways in which AI can be used in fashion design is a) for the entire creation process, or b) for assisting human designers in their creative process.

Although replacing human designers with algorithms across the entire sector is unlikely, there are certain areas in which designers are more vulnerable to the efficiency and data-driven approach of AI systems. Established brands that have a signature style and fast fashion brands that occupy themselves with quickly understanding and reproducing trends would greatly benefit from designing algorithms. This is because the focus is in producing new products with a similar style or design to what is already existent – the value of the item lies exclusively in the image, rather than the experience of engaging with the product, the history of its production and the creative process that led to it.

Stitch Fix[2] is an American online shopping service and the first brand to release a line created by algorithms. The company pins its success to the ability to understand and respond to trends, and to quantify how customers react to items that they see. The designs are a combination of online shopping data and customer feedback, and are a reflection of already existing trends and styles. In only relying data history, algorithms will continuously reproduce styles and trends that are already popular and well defined. This limitation leaves space for a creator or trendsetter, whose role it is to challenge trends and customer expectations and create a new product for people to respond to.

The example of Burberry’s recent re-branding is very revealing of this difference. If Amazon or Stitch Fix algorithms were to be used in Burberry, they would have been able to generate a number of items that fit within its traditional pattern. However, from looking at previous collections, it is very unlikely that AI would have been able to give the brand a new staple pattern as was designed by legendary art director Peter Savile. A creative can therefore generate the necessary discord to recode the design process, creating hype, press and fresh brand perceptions; whereas AI arguably won’t hit the necessarily disruptive patterns with such leaps of creative faith.

USING AI IN THE CREATIVE PROCESS

Using an AI systems to trigger inspiration during the creative process could lead to ground-breaking designs – if the system does not rely trending data, but focuses on the exploring combination of colours, materials and patterns, the designers would be less restricted by limitations of internalised prejudices and expectations of what is style and combinations. AI tools can unlock a new dimension in which designers can explore their trade, such as the use of virtual reality to explore new materials and the possibility of wearable technologies. Evidence of this potential was seen in the collaboration between the Fashion Institute of Technology and and Tommy Hilfiger[3], which used IBM’s Watson[4] AI system to analyse around 100,000 patterns from various fabric sources, producing unique and novel patterns from those. One of the final designs created was a plaid tech jacket, with which the student designer explored colour-changing fibres that respond to AI voice and social media analysis. Such systems better equip and inform designers about the possibilities of creation, and the experience resulted in very positive feedback[5]:

"As a fashion designer, I tend to stay in my own head, but with these tools, I was able to look into databases that were curated with an incredible amount of information, which, in turn, inspired in new ways that I could make design decisions faster." — Grace McCarty

"The traditional design process will always exist; for example, journal research, collecting fabrics, etc. But I believe that taking an innovative approach, like we did with this project, will help designers understand the new tools available to enhance our designs for the future. Technology is improving and changing by the day, and I think we are very lucky to have the chance to be a part of this movement." — Stefka de Ruiter

UNDERSTANDING THE PURPOSE OF FASHION

In spite of increasing efficiency and eliminating repetitive tasks, the use of AI in the creative asks us to consider the value of the creative process and the value of the personal experience of the designer. If inspiration is taken only from a database or the combination produced by AI – is the design (and the fashion) successful in serving its social function? One could argue that yes; current society is so intertwined with technology that excluding its input from the creative process would be denying a central shaping force in culture. On the other hand, the over-use of technology in fashion design can push aside the tradition of fashion and value derived from artisanal practice.

It is undeniable that the presence of AI systems will only increase in the future of fashion design, and to fully benefit from it in exploring creativity it is necessary to find a balance between experience and data. Fashion is, in its core, a sensory experience, intertwined with the way we humans experience the world and interact with each other. Algorithms are limited in accessing these experiences, but they can be very useful in a) understanding what customers want and expect, and b) prompting designers with new sources of inspiration that transcends their personal experience. It is not an either or scenario, and it is expected that both types of technologies will develop in tandem to each other. However, it is necessary to keep in mind the reason for why Fashion is valued, and why the creative and interrupting power of designers are very necessary.

[1] https://arxiv.org/pdf/1711.02231.pdf

[2] https://www.stitchfix.com/

[3] https://www.ibm.com/blogs/think/2018/01/tommyhilfiger-ai/

[4] https://www.ibm.com/watson/about/index.html

[5] http://www.businessinsider.fr/us/sc/ibm-fashion-fit-design-ai-2018-2