Why Fashion Forecasting will essentially remain ‘Human’ and AI will never ‘take over’

Last evening, I was conversing with an over-enthusiastic tech-addict startup founder who is in a ‘crusade’ to convince ‘the world’ that Artificial Intelligence is going to be the next ‘God’s particle’ and eventually human-brain has to forego ‘fashion forecasting’ process cause the “system can do it”. For him, the whole approach was more like a religion where one must ‘submit’ to the ‘almighty’ without brewing a ‘question’. Well, for many who argues that ‘AI’ is the last ‘Noah’s ark’, doubting it is a sin! The technology is indeed baffling when it comes to commerce. In a recent market report published by Markets and Markets on the topic of AI in fashion, they estimate that the size of the market is expected to grow from USD 228 million in 2019 to USD 1,260 million by 2024. From this forecasted period, it is believed to grow at a Compound Annual Growth Rate (CAGR) of 40.8%. Another report in Fortune Business Insights states that AI market size was USD 20.67 Billion in 2018. The growth and penetration of AI in the fashion market is significant. Mostly, to help shoppers find fashions that fit their bodies, budgets and styles, more retailers are turning to artificial intelligence. For instance, AI can help individual shoppers find the perfect pair of jeans by tracking their sales history, online browsing and Instagram likes for a seamless shopping experience. According to Intelligence Code, AI can also help consumers, who simply want “to buy stuff as quickly and easily as possible.” IBM writes that the adoption of AI-driven intelligent automation in the retail and consumer products industries is projected to leap from 40 percent of companies today to more than 80 percent in three years. In retail, the highest growth is expected in supply-chain planning, while in consumer products companies the deepest penetration is anticipated in manufacturing. Everybody is jumping in the bandwagon to part of new ‘opportunity’. The startups are mushrooming with a cohesive dream of getting ‘acquired’ by the giants (the ‘mid-summer dream’ of startup founders). This reminds me of those beautiful seasonal flowers (pun intended) craving to decorate someone’s living room, rather than evolving into a forest! The Analytics India Mag indicates that AI startups space is booming and now encompasses various avenues such as computer vision, self-driving, retail, audio production and innovating e-commerce platforms, among others. They also listed ten most ‘prospective’ AI startups from India. None of them were in Fashion Forecasting though. The three startups who claims that they are into ‘fashion forecasting’ ( I have been interacting personally for the past one year) are majorly into ‘data analysis’, ‘image recognition/ clustering’ and ‘retail sales analysis’. They are also into ‘heat mapping’ (digital/non-digital) and ‘trend insights’. One of them works in GAN and developed a system to ‘manipulate’ the existing images and combine them to create new images (for an e-commerce company), which they are claiming as a ‘new design creation’! They either don’t understand the Human-centric problem-solving approach of design (conventional design process) or thinks that ‘design-trick’ is as easy as downloading images from pinterest and converting them into a vector to claim as ‘unique design’. We cannot blame them as most of them are tech-addicts and hardly holds any intention to empathetically understand the domain of design or lacks intention to respectfully collaborate with any renowned design researcher/ team. When it comes to fashion, the state of knowledge is absolutely redundant beyond sales, manufacturing, and distribution. A quick Wiki read (as it’s the easiest reference for millennials to swear by it) itself states that “Fashion is a popular aesthetic expression in a certain time and context, especially in clothing, footwear, lifestyle, accessories, makeup, hairstyle and body proportions…Fashion instead describes the social and temporal system that "activates" dress as a social signifier in a certain time and context.” The understanding of “social signifier”, mass expression and attitude cannot be mapped quantitatively but to be understood with human emotion and empathy. The mega ‘zeitgeist’ (spirit of time) can only be mapped with human intelligence. However, the fragments of the trend as a marque of ‘change’ can be traced with a machine (when consumers stop buying green and browse more ‘blues’ in e-commerce site) which can also be done manually (through a strenuous time-taking method). It is crucial to understand the vast difference between ‘identifying trend’ and ‘forecasting a macro tendency’. Trend (in clothing/ makeup/accessories) can easily be identified with the system, especially if it is about the basic visual elements such as “the type of neckline” or “length of skirt”.

Hence, in future I hope that AI will enable us with ‘affordable’ tools to save time in data compilation. It should be as handy as a calculator, Tally or SPSS software for deriving cumulative results/ aggregating the details. The explanation of same and mapping the tendency (as insights) still remains a domain of the wonderful human brain. Of course, the insightful human brain can only explain why eventually “looking ugly becomes cool” and we spend a fortune on tattered denim/ oversized clothes as an expression of ‘self’ or to remain contemporary among the ‘peers’. 

Using AI, we can convert a still image into moving (with expression) but that again reflects our own imagination and pre-planned visual possibilities! The human brain still prevails as one of nature’s wonder. We are at a rudimentary technical level where technology yet can neither predict/control/stop bushfires in Australia nor can prevent a tsunami but only able to record it, map the quantum and faithfully propagate its visuals (of burning koalas) in social media for better ‘media engagement’ and ‘awareness’. In the book ‘Deep Thinking: Where Machine Intelligence Ends and Human Creativity Begins’ Gary Kasparov shows how humanity can rise to new heights with the help of our most extraordinary creations. He insists us to look into the future of intelligent machines and sees it bright with the possibility of working ‘together’ rather than fearing them of ‘taking over’.

(Article was written by Dr. Kaustav SenGupta on 23-01-2020. No portion to be printed/ published without permission)