The crop of fashion designers helming luxury brands are celebrities in their own right. They famously claim millions in a single contract but are equally worked and milked for their worth. To put things in context, some of them earn more than Joe Biden did in Barack Obama's persidential term. These are people who have 657,000 and 931,000 followers on social media platforms such as Instagram. Yet, why? Why are fashion designers celebrated and worshipped?
Since history, fashion has always been deemed the mirror of the zeitgeist. It reflects the socio-political moods of a time. Fashion has a sense of immediacy and instantaneous relation to the society that surrounds it. Fashion writer Jo-Ann Furniss once penned an article titled "The Shape of Things to Come". It highlighted the prophetic, creative nature of fashion designers. For decades past, the industry relied upon and championed this uncanny, prophetic human intuition.
What would happen if this very prophetic human intuition were translated into data, algorithms, and fed to self-learning Artificial Intelligence (AI) systems? Will AI help fashion designers create more precise and relevant clothes? Will AI help to resolve the recent spout of copying allegations, sizing issues, and help fashion houses read and respond to the ebbing consumer landscape? These are all, in fact, possible. There are numerous self-learning AI algorithms out there across varying creative sectors. The AI landscape has evolved greatly in the past decade.
Dr Vivienne Ming, theoretical neuroscientist and co-founder of the research institute Socos Labs, traces the history of modern AI back to December 1955, when Alan Newell and Herbert Simon successfully developed the Logic Theorist, a system that "played chess and solved math proofs." Fast forward to date, "the recent rise of AI has its roots in what is called 'machine learning' — computer programmes that master a task by learning from experience." A defining characteristic of AI is, therefore, its ability to self-learn. When AI systems are trained adequately, they can outperform humans — and open new doors to the fashion industry.
AI in the Fashion Industry
AI has been introduced to the fashion industry on numerous occasions to optimise the design, manufacturing and retail decision-making processes.
Earlier this year in January, Forbes reported that the American brand Tommy Hilfiger, The Fashion Institute of Technology and IBM came together to explore the potential of AI in fashion design. Data from the brand's archives were processed and analysed. "The machine learning analysis gave us insights about the Tommy Hilfiger colours, silhouettes and prints that we couldn't begin to consume or understand with the human mind," a representative concluded. Take, for instance, the key designs and styles of a brand, the consumers' reception to past and new designs, and the ongoing fashion trends in various geographical contexts can inform the design team in their work.
For multi-label stores and brand boutiques, AI can prove to help predict consumer behaviour and highlight which designs will potentially be well-received. In 2011, it was reported that New York-based fashion designer Elie Tahari introduced an IBM data analytics tool, one that optimises the manufacturing arm of the business. The system essentially optimises a merchandiser's role by crunching data and generating a report of the most (and least) popular designs, colourways, and sizes.
When it comes to the retailing of clothes, numerous brands have hopped on the bandwagon of personalisation softwares for consumers. Yet, not all of them are considered AI systems. The majority of them are merely automated systems. A similar software service, Savitude gathers data about the user's body shape before self-learning algorithms recommend clothes that might fit their body proportion. It proposes to resolve the sizing issues that consumers often face while shopping online.
Other issues that AI could potentially resolve includes the recent spout of fashion copying for instance. There have been numerous allegations of high street brands such as Zara and H&M copying runway fashion. There also have been allegations of established designers copying from the younger ones. The situation has prompted fashion insiders to question, 'Why is there not a plagiarism detection system for fashion?'
An academic plagiarism detection system like Urkund, the direct rival of TurnItIn, is essentially an AI system. "We are using a lot of methods that are considered to be AI — machine learning and so on," Peter Witasp, the chief operating officer at Urkund explains. Students basically turn in their academic drafts to the online system, and the algorithms analyse many aspects of the essay. "Some of the steps involved are text extraction, text indexing, stylometry indexing, semantical indexing, comparison, analysis, text alignment and report generation. The system is self-learning in several senses."
A plagiarism report generated by Urkund.
Urkund came about in the '90s, when "the internet was introduced at universities as a research tool". The internet made it easier for students to copy and paste texts that the found online. A few teachers at the Uppsala University in Sweden noticed the trend. They manually searched up these dubious texts on search engines to look for their sources. It was eventually made into an automated system. Likewise in the fashion industry, the internet has arguably made copying and referencing easier. Is a similar AI-driven plagiarism detection possible in creative industries? "Yes, sure, it is where the world is going," Witasp responds.
There are numerous AI visual and object recognition systems on the market at the moment. A fashion plagiarism detection system will likely serve as a reference tool for fashion brands. How it might potentially work is this — the designs have to be submitted in a standard visual format, and it will be compared to the bank of designs that presumably have already been uploaded into the database. The AI system will be able to learn to recognise the style and fingerprint of various designers. Later, a similarity report will be generated.
A tool like this will serve as an insurance of designers' livelihoods. "When it comes to the creative industry, the people who are the creators, it is their livelihood... If someone is stealing your pattern, you as the creator will not get the recognition and live off your work," Witasp continues.
Be it plagiarism detection or predictive analysis, AI systems are merely reference tools informing newness and novelty. Will some merchandisers and trend analysts be displaced by AI? Likely. Will fashion designers be displaced by creative AI algorithms? Unlikely. The truth of AI applications is this — they are tools to bypass routine tasks. "In the end, the human really is the one being creative," an article published on IBM's website observes. The prowess to respond to environmental stimuli like a social or political upheaval remains very much intuitive.
A trade report revealed that AI will potentially dominate the fashion industry in the year to come. How will the introduction of AI change the industry? It will likely propel brands to produce more relevant and precise designs for consumers of different shapes and sizes. It will potentially speed up the design, manufacture and retail processes — hurtling the rapid-fire industry into an even faster roulette.
Subscribe to our newsletter