COVID-19 lockdowns have wrought catastrophic effects on brick-and-mortar retail worldwide. With the increasing digitisation of fashion, I explore how A.I. is revolutionising the way we shop and create.
You. Yes, you. Have you ever paused to consider that you’re one of the world’s most precious resources? You and most everybody around you are walking mines of something all businesses want: data.
Companies need to know how to sell to you, and every impression you leave behind while window-shopping online, catching up on the latest celebrity news, or liking a photo on social media helps them achieve this. Like a unit of poachers who obsess over animal tracks, they’re constantly watching your movements and trying to get one step ahead, by building as detailed an outline as possible of your preferences and habits. While you’re hunting for the best bargains, retailers are hunting you.
Pinning you down is a far more complex process than you might imagine. It’s often a team effort, with online platforms trading the dirt they have on you under the pretext of delivering a more seamless and customised experience. Logging into The Business of Fashion or Spotify using your Facebook account, for example, means these platforms gain access to your existing FB data and don’t have to profile you from scratch. If you’ve ever wondered how the internet seems to know you inside out, think of it this way: ASOS, YouTube and Instagram have not only been hanging out while you’ve been asleep or offline… You’ve also been their main topic of discussion.
Which brings us to deep learning, currently the hottest field in artificial intelligence. Those movie robots that learn and grow increasingly sophisticated have become reality. While they look nothing like Alicia Vikander in Ex Machina or Jude Law in A.I.(unfortunately for us, amirite?) they are very, very good at predicting what we’ll like. Algorithms like those controlling your Instagram newsfeed are continuously, independently learning from their ‘mistakes’ and self-correcting, in the same way a flesh-and-blood person would. Their ultimate goal is getting as much of your time and money as possible.
With human civilisation producing more text and images in the last thirty years than in the past two millennia, much of which floats freely on the web, A.I. systems have plenty of material to study us with. It’s worth considering, though… Might artificial intelligence someday cross the line from being proactive to being intrusive, or from being helpful to being despotic? Will robots and computer programs overrule us, exerting a disproportionate influence on what we wear or deem cool? Whatever the answers, one thing is for certain. These concerns are not as far off in the future as we’d like to think.
During the early 20th century, when research in A.I. was at its most wildly fertile, computers were literally and fittingly like babies. They were chunky, less adorably so than human infants. While they demonstrated potential, they required a ton of supervision and input, and were incapable of self-maintenance or doing anything they weren’t explicitly told.
How things have changed. Like awkward child stars who blossom into beauty and become fodder for thirsty memes, computers glowed up and, these days, enjoy previously unimaginable levels of ubiquity. They’re now compact enough to be taken everywhere, and are more intelligent and less awkward than ever. Their friend list has exploded. Even inhabitants of some of earth’s most remote locations own smartphones and complain regularly of crappy network service. Like a Sk8r Boi-turned-rockstar, A.I. may soon realise that it’s too good for us.
The technology that once desired nothing more than to listen and be of use has developed increasingly controlling tendencies. Before you accuse us of being dramatic (which we’ll half-cop to) let us remind you of Miranda Priestly’s cerulean belt monologue from The Devil Wears Prada, which summed up the power of trends to reverberate far beyond their point of origin. Somebody is indirectly responsible for the textures of the world around you, from bra colourways at Lululemon to the ergonomics of iPhones. Whether trends bubble up from the streets or trickle down from the high-fashion runways, there is always a voice driving the movement. That voice, in today’s landscape, is progressively more digital.
Trend forecasting and analytics are big business. While humans still work behind the scenes of your favourite brands, retailers, and content platforms, A.I. is having a larger say in defining the zeitgeist. Its numbers don’t lie, or so we’ve been led to believe. Not drinking the A.I. Koolaid is asking for a serious case of economic FOMO. Many fashion companies, big and small, now live — and die — by the “I saw Cady Heron wearing army pants and flip-flops, so I bought army pants and flip-flops.” philosophy.
Don’t believe us? Take a look back at Alessandro Michele’s kooky, fall 2015 Gucci debut. In the days following his first runway show, the establishment’s reaction was tentatively supportive, unprepared as they were for a massive shift in direction. The larger audience, on the other hand, was starved for quirk and humour. They made their feelings known on social media, and that was that. Data was crunched, and a torrent of Gucci-inspired, “daft” fur-lined slides swiftly flooded stores of all market levels. Michele’s victory over the industry’s conservative gatekeepers has since been recounted as a triumphant underdog’s tale. It also raised the troubling prospect that modern culture might be nothing more than a numbers game. A digital, figure-based assessment might soon become the only metric by which creative success is measured.
This wouldn’t be a problem if A.I. just told us what was going on in real-time, as it did a few years ago. However, as we spend more of our lives online and the internet’s web of connections grows more complex, the role of artificial intelligence is no longer reactive, but prophetic. Everyone, and we mean everyone, is transfixed by its pronouncements, though few are stopping to ask if blind faith in A.I. diktats is making the fashion landscape more homogeneous, and consequently, more boring. Will A.I. prematurely kill ‘fetch’ and make compliant Plastics of us all? Will shopping cease to be fun when every boutique sells nothing except f*cking fur slides?
DESIGNING WITH DATA
Fashion has traditionally been terrible at user experience. Don’t @ us. For decades, designers have sent imaginative clothes down the runway which are irrelevant in the wardrobes of most ordinary people. This disconnect between aspirational and practical design is not only frustrating for us shoppers, who mostly wait to buy the watered down, more wearable (and, let’s be real, more economical too, because millennial paychecks ain’t sh*t) versions of designer pieces. It could also be financially disastrous for creativity-driven, high-fashion businesses.
In large part thanks to the internet and social media, the rag trade is now global, decentralised, and saturated with labels of all sorts and sizes. This is fantastic for consumers like us, who are more label-promiscuous (no judgement, obvi) and eclectic in our tastes than ever. The socialite Mona Von Bismarck famously went into mourning when Cristobal Balenciaga retired in 1968. Meanwhile, our favourite label could disappear tomorrow and we’d likely have a million alternatives at the ready.
This hyper-competitive state of affairs may not revolutionise fashion design as much as ruin it. You see, it takes time and money to develop original work, which runs the risk of bombing on the sales floor despite critical popularity. Alexander McQueen the label, for example, only broke even after fifteen years in business. Every dollar counts for fashion brands in 2020, and few can afford refusing A.I.’s trend forecasting arm in maximising profit margins.
For instance, social media analytics provide feedback on what we consumers are visually and conceptually drawn to, which supports image building. Most of us may not buy into a full-on Gucci floral frock with 80s-esque volume. We may, though, be drawn enough to the idea of it that purchasing a sensible shirt in the same pattern carries an equivalent thrill. It’s sales analytics that reveal what makes ultimately money and what doesn’t. Search result clicks, add-to-cart wishlists (we know you have one) and abandoned checkouts all tell designers a story of where the perfect balance between want and need — or like and buy — resides.
Let’s face facts. For the most part, we shoppers… are basic. Our tastes are sooo basic, you guys. Handed the wheel to steer the fashion industry with our choices, we probably wouldn’t take it anywhere new or interesting. That’s what designers are there for, to remind us of clothing’s cultural, ludic, and less mundane functions.
Fashion, of course, should respond to people’s day-to-day needs. Is that all, however, that it is meant to do? It’s wonderful that we clothing-wearers finally have recourse when brands— many of which have historically held nasty, exclusionary outlooks — neglect our identities, our body shapes and our less outlandish tastes. However, one of the reasons why many of us love fashion is that, at its best, it can provoke and challenge us.
Designers like the late Alexander McQueen and Viktor & Rolf — in their zany, let’s-pile-20-layers-on-some-hapless-model prime — occasionally elevated fashion to sublime heights. Their shows weren’t just about the key silhouettes and accessories to wear next season. Those were evident, sure, but attendees and browsers on Style.com (RIP) also got a narrative breadcrumb trail hinting at bigger themes and ideas unfolding in the world.
McQueen’s Plato’s Atlantis show of spring 2010, for example, was more than a parade of reptilian, immaculately tailored gazar mini-dresses. It was also more than the debut of Lady Gaga’s megahit Bad Romance, outsized turnout for which crashed the event’s live-stream. Each otherworldly look formed part of a fable spun by the designer, about climate change, a drowned earth, and a human race that evolved into aquatic creatures in response. Sure, a nutty (and scientifically shaky) fantasy like that isn’t, strictly speaking, necessary for bringing a series of pretty frocks into the world. Although golly, isn’t it dreamy when it happens that way? In the words of fashion dowager Diana Vreeland, great fashion should “Give ’em what they never knew they wanted.”
A.I. has made it impossible to ignore the voice of the ordinary consumer, hitherto fashion’s ‘silent majority’. Having said that: is giving financially-pressurised designers reliable, practically unquestionable insights on ROI (return on investment) inadvertently reining in their creative impulses?
In 2018, designer Zac Posen told Observer: “A.I. will not be able to synthesize the irrational surprises that only humans have… situational, spontaneous moments of beauty (disappear) when things are formulated and automated,” So… finger-snapping, sassy robots probably aren’t coming for Posen’s job anytime soon. But the industry should ask itself if A.I. is revolutionising the way designers serve us in the short-term, while ruining fashion in the long -run by depriving us of their full creative intensity.
FASHION LABELS: THE BAD KIND, AND HOW A.I. CAN DIVIDE US
“Do we all see colours the same?” sounds like the kind of line some pretentious, tipsy university student might blurt out at a house party. Nevertheless, an interesting conceit lies buried within that douche-y query. Person to person, our experiences of the world may not be equal.
Online and in real life, social divides are growing wider. The data-based, highly personalised internet bubbles that Facebook, Twitter and other web giants tailor for us may be to blame. See, our digital footprint not only helps them to profile our wallets and attention spans, but those of our demographic peers — that is, people of the same gender, race, age group, income bracket, etc. — as well. Google calls this demographic targeting, and it’s controversial AF.
Netflix, for one, has come under fire for allegedly using algorithms to adjust recommended programs’ artwork for each viewer. For instance, they’ve been accused of advertising American films with ethnic-minority leads to white people using imagery of white minor characters. Such choices may be based on the reductive and dangerous reasoning that white people can’t — and perhaps shouldn’t have to — relate to people of other races. Similar systems could and may already be in effect on social media, narrowing platforms like our Instagram Explore pages to exclusively display content from people who resemble us. A.I. may not be capable of understanding and holding prejudices of its own, although it sure as hell can help us act on ours.
Whether or not demographic targeting is an efficient marketing strategy — and it probably is, given how much money businesses continue to pump into it — it’s hard to deny that it fuels our conscious and unconscious biases. In the pursuit of bottom lines, powerful companies demonstrate little hesitation in exploiting our ignorance. Only a decade or so into the Social Media Age, we have been filtered into what are at best bias-enforcing bubbles and, at worst, cultural ghettos.
Demographic targeting has yet to affect fashion much, however that’s only because the industry is, as usual, fashionably late to the tech party. Nonetheless, it’s hardly a stretch to imagine a near-future in which fashion’s bigoted big players (salty side-eye at the industry’s Ed Razeks and Stefano Gabbanas) use deep learning algorithms to identify and quietly exclude entire classes of ‘undesirable’ people from their advertising budgets. Those controlling the means to advertise, meanwhile, could in theory stratify audiences, then charge premium rates to reach those considered more ‘desirable’.
Even seemingly harmless AI devices such as Amazon’s Echo Look (now discontinued) could replicate ugly prejudices. The style-assistant camera relied on a combination of deep learning and human input. Through user-generated OOTDs and crowd-sourced opinions thereof, it aimed to assess users’ outfits, select the ‘best’ look from a series of uploaded pictures, and offer shopping recommendations from Amazon’s vast inventory. Echo Look was constantly adapting by processing massive volumes of the latest fashion imagery and user feedback. Given the very flawed state of contemporary fashion, though, could basing (if only due to the current limits of technology) the output of devices like Echo Look on existing data be a bad thing?
At the moment, A.I. can anticipate trends from what is presently visible in the world. As Posen pointed out, it has yet to develop enough sophistication to simulate genuine originality or brilliance. You know those annoying, TV makeover stylists who parrot every fashion “rule” that’s ever been published in a glossy magazine? That’s pretty much what Echo Look and its surviving relatives are. For now, A.I. has no capacity to understand the uniqueness and expressive potential of each human being (and unique and expressive we are, even if we don’t send snake-women slithering down a catwalk ourselves). Although there’s no shame in seeking help in the style department, we might pay a heavy price for digitising how we get dressed.
Let’s not mince words. Fashion has typically been sh*tty to larger (whatever that means to you) women. While magazines and websites offer those who fit the mainstream definition of ‘average-sized’ endless new ideas to try, ‘plus-sized’ women are more often told what to avoid or minimise. It is a deeply unfair, depressing state of affairs. Despite the recent rise of independent and size-inclusive clothing labels, these disruptors remain as yet a minority.
So, if Echo Look-style A.I. systems are making decisions or learning based solely on available data, guess what? Were it assigned a human personality, it would probably be a small-minded Ed Razek instead of a progressive Tan France. A.I. has ostensibly revolutionised personal style, empowering us to make better fashion choices. But does ‘better’ in this instance simply mean more conformist or ‘tasteful’, and less expressive or risktaking?
ARE WE LIVING IN THE AGE OF THE DIGITAL SWEATSHOP?
None of us like finding out that the clothes we wear are made of human suffering. However, it’s time we had a serious ethical discussion about how technology and A.I. have redefined the value of human labour, especially in the fashion industry.
“Reach, not preach!” you cry, “Garment workers have always been underpaid and ill-treated! What’s A.I. got to do with this?” As it turns out, lots. These days, the term luddite describes those resistant to industrialisation and technological progress. It originally referred to mobs of predominantly textile-based labourers, who rioted and destroyed machinery in the emerging factories of 19th century England. While there was no halting the march of progress, the Luddites had pretty compelling reasons for their smash-happiness.
Mechanical looms and sewing machines revolutionised the way clothes were made, no question. But while they streamlined the process, things didn’t necessarily get better for the people behind the equipment. When part of a worker’s job is automated, they’re unlikely to be paid for it, even if a mastery of the entire process remains essential on their part. The advent of the sewing machine didn’t mean that seamstresses no longer required knowledge to sew. We’ve come a long way from the catastrophic upheavals (at least for workers) of the Industrial Revolution, but similar developments now threaten the livelihoods of those on the lower rungs of fashion’s manufacturing chain.
Just as machine-sewing supplanted hand-stitching, some hitherto manual design processes are being taken over, at least partially, by artificial intelligence. Software for pattern grading (taking a sample-sized clothing pattern and scaling it up and down for different sizes) and lay planning (figuring out the minimum-waste cutting layouts for pattern pieces on fabric) is already widely available. They save plenty of time, materials and effort, but arguably devalue human skill.
That devaluation isn’t indiscriminate, with several writers and tech observers documenting the ‘artisan economy’ that has arisen from automation and A.I.’s disruptive influence. In a market where the machine-made predominates, handcrafted or human-intensive work in its most rarefied forms can command premium prices. This has led to the polarisation of the fashion industry that we see today, where designer fashion is, adjusted for inflation, far more expensive than it’s ever been (dammit), and factory-made clothes from developing countries retail for as little as a handful of dollars *fist-pumps, then feels extremely guilty*.
Though the middle ground between these two extremes hasn’t completely vanished, most of the money to be had in fashion now goes either to luxury mega-houses or to the ridiculously cheap leviathans of the mass market. Very little of these huge revenues is making its way into the pockets of workers. At the same time, the paychecks of CEOs, star designers and image-builders have been shielded from clothing’s overall diminishing value. In fact, they’ve become billionaires and rockstars in a way even veritable fashion gods like Coco Chanel or Christian Dior would have found difficult to fathom. The fashion industry and its workings hold a mirror to society, and the picture it’s currently painting is extremely grim. Factory workers remain in poverty, the rich fashion folk get richer, and the middle class of workers find themselves increasingly pressed to the margins.
Old notions that A.I. could only ever make repetitive, unimaginative work obsolete have been shattered. It’s depressing to think that competing against artificial intelligence has reduced us to little more than machines. To keep up, we have to be as inhuman as possible, else we be penalised for natural phenomena like aging, illness, or just plain tiredness. All this begs the question: is there any hope for an ethical fashion industry?
Countless works of fiction imagine idyllic worlds in which all drudgery and exploitation have been banished. Whether A.I. is the innovation that will carry us to those idylls, or how we’ll get there from today’s work-or-die capitalist societies, we have yet to figure out. Proposals like a four-day work week and universal minimum wage are still remote, despite gaining momentum during the COVID-19 pandemic.
We must question if technological advancements as we pursue them are equalisers, or if they instead improve the lives of the few to the ruinous detriment of the many. If the A.I. revolution is transforming the way we dress and adopt trends, but leaving millions excluded and exploited… There may be nothing ‘revolutionary’ about it after all.
This story was originally featured in the broadsheet Buro. Large.
If you’ve enjoyed my work and would like to support my nascent, independent writing/research (a girl has bills to pay, I’m sure y’all can relate), please consider donating to my Patreon. While Patreon is based on recurring payments, you can unsubscribe at any time, which would basically be like making a finite contribution! This bashful broke bitch thanks you in advance.
You can also follow me on Instagram (@theryaninwinter). Feedback on my work is welcome, but be warned that I’m extremely thin-skinned and delicate, gourmet soup dumpling-style. I’m not quite as bad as those “pOsiTIVe viBes OnLy” folks… although we’ll get there eventually, I’m sure of it ;).
