About Fashion, Identity and Artificial Intelligence

I’m quite far from being a fashion addict. It’s undeniable that clothes are one of the way we do project ourselves into the world. Clothes should be a physical embodiment of how we portray ourselves…but I never found myself represented by one style, brand or garment. I always had the feeling I was wearing a costume, some kind of disguise. A costume for a job or for a party. Or a costume as a status symbol. Therefore, my style is something that would be defined as normcore. Not representing a true me, not personal, just standard clothes to adapt to society requests. 

My relationship with AI, instead, would be labelled, as “It’s complicated”. First time we met, it was a rainy afternoon, November 2000, at Bologna University. Like in a quintessential love story, I bumped, by chance, into a seminar on Artificial Intelligence. Aha ah moment. To quote Andrew Ng, I also thought that making machines intelligent was the coolest thing that someone could ever do. Five years later, despite AI was considered an exotic discipline, I managed to write a thesis on Semantic Search Engines and almost start a PhD. However, reality took over and I (happily) ended up in a product management job for Skype. 

Myself and AI – and allow me to use AI as Augmented Intelligence…Artificial implies just a cheap imitation of the genuine article. Distrust and resentment built into the word itself. So myself and AI, after University, kept meeting each other time to time, as lovers, not as a stable couple. Few years ago, machine learning hit the news, I started being involved in AI in Oracle and teaching Future of Fashion at Polimoda. Ah aha moment again: Fashion and AI. Connecting the dots, eventually.

Let’s see some numbers. The value of the global fashion industry is US $3T, 2% of the world’s GDP.

3B textile and garment new companies are entering in the market daily, but most of us are influenced only by the big giants like LVMH, currently the largest fashion company worldwide, Nike the second largest and fast fashion king Inditex is at the third position.

 Estimating the market value for AI is a bit more complex since it’s a horizontal technology that unlocks a potential market value for several other industries. McKinsey estimates that AI and machine learning will create, by 2020, an additional $2.6T in Marketing and Sales, and up to $2T in manufacturing and supply chain, most of it in Retail.

Both Fashion and AI are constantly under the spotlight, often mystified. They both suffer from overstatement. The main myth?

Fashion Designers are creative whilst AI cannot be creative.

Let’s start with the first part of the myth: Fashion Designers are creative.

How does a Fashion Designer create a new collection?

First, she/he has to follow the brand style guidelines. Second trends forecasting are determining her decision-making about textiles, materials, color palette. Businesses crave predictability. 

If you’re wearing today an ultraviolet garment, it’s because a designer gave credit to a prediction issued by a trend forecaster at least 2 years ago. And more than 12,000 brands get the trends forecasting from the same agency. Trends Predictions are, in reality, Trends Creation. Third: a designer could simply pick up one bomber jacket from 1943 US army and launch it, exactly as it is, as her new collection. In other industries, this would be considered plagiarism. In Fashion, this designer could be regarded as a genius. Apparel is classified as “utilitarian”, so usually trademark is protected, rarely copyrighted. Look at fast fashion. It’s often a copycat of high-end design, done with cheaper materials. 

Many mundane activities are already done with the help of technology. Fashion designers don’t sketch anymore. They use 3D software to design a dress and to modify it in real time. 

The key enabler for the Digitalization of the fashion industry are data. Zara vertical integration is a legendary money machine. Data is the magic wand to predict both demand and supply. Zara pioneered the usage of PDAs (personal digital assistants) to capture real-time consumers’ data for both transactions and preferences, connecting then consumers ‘data not only to marketing and communication but also to production and supply chain. To eliminate months of lead-time, Zara design teams are given a limited set of pre-selected fabrics and materials they can use. Predictability already won over creative freedom. 

And what about the second myth?

 AI cannot be creative.

https://www.youtube.com/watch?v=4MKAf6YX_7M

The music you can listen playing the video was created by the Israeli production team called Sweaty Machines, a team made by humans like Israeli poet and programmer Eran Hadas and machines like Oracle AI. The algorithm analyzed hundreds of Eurovision songs written by humans to come up with sentences like 'There’s no life without your life in misery' or "Tears will always have wet eyes.” The song, titled “Blue Jeans and Bloody tears” aspired to become a hit – a sort of unofficial anthem - for this year’s Eurovision contest. 

In April "DeepVogue" AI, developed by the Chinese DeepBlue Technology, won the 2° place at the China International Fashion Design Innovation Competition beating out 16 human counterparts in front of a panel of 50 judges. 

Image: Design created by Deep Vogue

Twitter users reacted enthusiastically: “it looks better than many of FW human creations”. 

In November 2018, Yoox launched 8 by Yoox, a collection created through algorithms that draws trends from a plethora of data source. It produces a mood board and later a collection of essentials. It’s probably much more accurate in predicting what people will truly buy and wear than a fashion designer.

Image: 8 by Yoox

And what about using AI to boost your mood? Studies confirmed that emotions are influencing what we do wear and what we do buy. A study by Professor Karen Pine, from the University of Hertfordshire, found that 57% of women wear jeans when they have a blue day. "Happy" clothes tend to be those that are flattering, well-tailored and made from bright fabrics.

Does it mean that fashion designers should look for another job?

Not at all. The bond between algorithms and data doesn’t mean the death of true creativity. Quite the opposite. As shown by Robert Barat, who trained an algorithm to create Balenciaga garments, what deep learning can do is to make a good copy or good variations. Machine learning finds patterns in data, numbers or images, and replicates it.

Is AI creative or not? To quote the computer scientist E. Dijkstra:

“The question of whether machines can think is about as relevant as the question of whether submarines can swim”.

If AI is creative or not, is simply not the right question. AI can produce a design we’d love to wear. Data enables to take into account the context of individual emotions, patterns, and experiences. Machine learning is already helping companies to move from mass marketing to hyper-personalized communication, suggesting the best time to send out an offer or how to craft the most effective message or what is the best channel to reach you.

Robotic Process Automation, AI, machine learning, NLP, robotics allow us to automate many boring and mundane activities. Have you ever been in the “flow”? That state when you are completely absorbed in what you do, so absorbed that you lose the sense of time? That state when you get creative epiphanies or eureka moments? I bet you’ve been there, but not so often. You’re too busy with commuting to work, rushing to meetings, responding to texts and emails, performing other mundane activities. What if you could delegate all those activities to an AI system, able also to schedule your day so that you weren’t constantly interrupted and knocked out of the zone? AI may allow us to have more time to think, to feel, to create. To have more time to be human. Smarter Humans.

The right questions to ask should then be:

How can we be sure that we are not developing technology for the sake of technology? 

How could we embrace AI in order to make humans smarter?

By co-designing with AI. By embracing what Professor Ken Golberg calls Multiplicity. Diverse combinations of people and machines working together to solve problems and innovate. Humans are not the fastest or the strongest species, but we are the best learners. We should learn that the evolution of our creativity sits in the crosshairs of art, science, engineering, and design. 

Few months ago, the world’s first piece of digital couture created by The Fabricant was sold for $9,500. It’s validated by the blockchain – it’s both a garment and a cryptocurrency. The new owner has a 28-day window to provide a photo to the human creators in order for them to custom fit the digital garments. Humans, technology and machines together to crave innovation.

Stich fix is already creating “Hybrid Design” garments. The algorithms identify trends and styles missing from the Stitch Fix inventory and suggest new designs — based on combinations of consumers’ favorite colors, patterns, and textiles. Suggestions that should be reviewed adjusted and approved by a human designer.

We are entering the AI driven

hyper-personalization era.

What does it mean?

That you are part of the game. You will influence the fashion trends. Your preferences, your personality your habits, your lifestyle will allow a team of Human Designers, algorithms and machines, to propose you a unique design, a tailor-made size, a temperature sensitive fabric adjusting specifically to you. To you, an individual with a unique set of exceptional characteristics and a unique personality. No more normcore clothes just to address standard social requirements. Bespoke, hyper-personalised garnments matching your true self or maybe the self you are just becoming. For all of us. No prohibitive pricing or long body measurement session. Your dynamic avatar will provide to the designers the right measures.

No more this feeling of wearing a costume, of being in disguise as someone else, not loyal to yourself.

If Love is about trust and acknowledgement then hyper-personalised clothes will be the way not only to project yourself into the world but the best way to truly love yourself.

Luisella Giani