Cin7

cin7 logo

Blog

How Fashion Tries AI on For Size

by Simon Eskow
how-fashion-tries-ai-on-for-size-jpg
Tech entrepreneur Elon Musk worries about the threat AI poses to humanity. Speaking to a group of US governors last weekend, Musk made an atypical appeal for government regulation. His fear is the uncontrolled development of artificial intelligence.

“AI is a fundamental risk to the existence of human civilization, in a way that car accidents, airplane crashes, faulty drugs, or bad food were not.”

First, unsurprisingly, artificial intelligence will threaten jobs. However, that is a perennial side-effect of every technological innovation since the dawn of the Industrial Revolution. Then, as Fortune summarized it somewhat cryptically, it may lead to social destabilization. Finally, it can lead to war.

Musk posed a scenario in which artificial intelligence programmed to maximize profit could, theoretically, launch a war to manipulate markets.

Is AI All the Rage in Fashion?

What does any of this have to do with fashion? For starters, the fashion industry has proven a fertile petri dish for developing practical applications of AI. Researchers show how machine learning applied to big data and image analysis can lead to deeper fashion insights. The end result of fashion AI is to help brands create supply based on buyer preferences. But as everyone knows, there’s no accounting for taste.

Still, AI seems to be getting brands closer to that goal. For example, we just read this week that Hybrid Designs, the house brand of fashion subscription box provider Stitch Fix, employs data scientists to examine customer orders to plug holes in its inventory. The team analyzes orders to come up with 30 trillion potential clothing item combinations to come up with a small number of suggestions for the brand’s design team to create a new and unique design concept.

What’s the Current threat level? Fashion Faux Pas

Musk’s dire warnings may be premature, going by examples of fashion AI. Collectively speaking, fashion AI has yet to really achieve the goal of matching supply to customer demand. A case in point was the early iteration of Google’s Project Muze.

In 2016, Google teamed up with a European eCommerce company for Project Muze. In this experiment, Google’s neural network took fashion tips from 600 experts. They taught it about styles, color and fabrics. Then they told it to come up with creative ideas for new fashion concepts.

The results weren’t ideal, prompting TechCrunch to give it a humorous chiding.