Great ResignationClimate ChangeLeadershipInflationUkraine Invasion

Commentary: AI’s Next Victim: Your Closet

March 15, 2018, 1:00 PM UTC

In 1995’s Clueless, you may recall Cher Horowitz using cutting-edge software to select her plaid ensemble. Cher’s machine could identify chic head-to-toe looks, adding a small dose of sci-fi to the romcom classic. Twenty-two years later, the 90s fiction movie is closer than ever to reality: Artificial intelligence in fashion is here, but it’s still unclear what its role is meant to be.

As a form of personal expression, fashion may seem like a strange target for AI disruption. Regardless, machine learning is taking on a variety of fashion-related roles. These include personal shopper, fashion designer, model, and—just like Cher’s program—stylist.

Case in point is Pureple, an app that provides a similar service. After prompting you to photograph and categorize items in your wardrobe, the app suggests outfits that you can “swipe” left or right on to refine your preferences. How well does it work? YouTube blogger Safiya Nygaard took its suggestions at face value for a week and was moderately impressed with its choices in spite of what she described as tedious upkeep.

On average, people only wear about 20% of their wardrobe. AI like Pureple, then, could help utilize those long-ignored items by suggesting totally new outfits. AI has a tendency toward unusual combinations: IBM Watson’s AI-generated recipes, for instance, combine flavors that human chefs instinctively keep separate. Imagine the outfit version of a Swiss-Thai asparagus quiche. Would you wear that? Maybe! Would you come up with it yourself? Probably not.

Apps like Pureple are exciting, though their usefulness remains up in the air. Similar innovations are emerging. Kim Kardashian West-backed Screenshop, for instance, finds items you can purchase using AI image recognition. Like a Soundhound for clothes, such a product could make shopping more convenient, but seems unlikely to usurp the in-store browsing experience.

On a grander scale, it should be no surprise that tech behemoth Amazon has begun using AI to design clothes. According to MIT Technology Review, an Amazon research team “has developed an algorithm that learns about a particular style of fashion from images, and can then generate new items in similar styles from scratch.” This is in addition to an existing app, Echo Look, that provides feedback on your outfit selections.

Shoppers may still prefer to support human designers, and have more faith in their abilities. But with AI technology still emerging, we could see AI-developed fashion lines in stores before we know it.

To bring things back to Clueless, have you ever wondered how an item of clothing would look on whoever you’re buying it for? Cher’s program clothed an avatar that looked just like her. Similarly, the startup’s technology is doing the same for brands and retailers. According to Quartz, the product “analyzes pieces of clothing and automatically generates an image of the garment on a person of any size, shape, or wearing any kind of shoes.” In an industry that still works with mostly waifish models, this could be a game changer.

It’s clear that tech is smarter than ever in the fashion industry. But intelligence will mean nothing unless these innovations augment the personal and creative elements of fashion—instead of completely automating them. Fortunately, I think that’s what’s likely to happen as these technologies continue to mature.

Devorah Rose is the editor-in-chief of Social Life. She has no investments in or associations with the companies mentioned in this article.