Get Started

AI in Fashion: Fad, or the Future?

Use Cases & Projects Rose Wijnberg

AI has been part of the equation in fashion for years — if you’ve shopped online, you’ve probably experienced AI-powered styling (whether you knew it or not). We’ve already delved into the key applications of machine learning in the fashion industry, but use cases like recommendation engines are just the tip of the iceberg when it comes to the role AI can play in the fashion industry.

→ Get the Ebook: Top Trends & Opportunities in AI for Retail

AI & Us, the new web series from Dataiku, explores how AI is changing our everyday lives, starting with what we wear. Given that fashion is one of the biggest industries in the world, and that the lifespan of the average garment is just five weeks, this inaugural episode asks big questions. Like, can AI be creative? Start trends? Or perhaps more importantly today, can AI help make the fashion industry more sustainable, inclusive, and even accessible? We — and our experts — don't have all the answers, but asking the right questions is a start.

 

Big questions aside, across all industries (including in fashion) there are two key best practices that will help ensure a more positive AI future.

Human-in-the-Loop Design

Whatever the task — from creating trends, for AI in fashion, to other industry use cases such as predicting loan delinquency, determining insurance premiums or payouts, preventing fraudulent transactions, analyzing customer feedback, forecasting energy use — automation is part of the story, but it’s not the entire story.

Yes, machines might be better suited to certain tasks, especially repetitive jobs. They don't get bored, they deliver consistent quality, and they don't leave or demand a transfer to a better paying, more exciting job. However, we need people and automation. We need humans and machines each doing their best work. With humans in the loop, people can work faster, but also make sure our AI systems align with human values, do what we expect, and that we can continue to trust their outputs.

Tools that Put AI in the Hands of the Many

For AI to be a success, no matter the industry, it needs to be accessible to those with deep business knowledge, not siloed to only data scientists or other data-centric roles. Accessible means not only in terms of leveraging AI, but building it using tools such as no-code or low-code solutions, as well as being able to understand its outputs. 

In other words, being able to say “so the machine predicted X… why? And what does that mean? What is the outcome?”

The web series AI & Us will explore these best practices as applied to various industries and challenges, like insurance and the gender pay gap. It will also delve into our perceptions about AI itself and whether we’re heading in the right direction when it comes to trust, transparency, and more.

You May Also Like

Graph Neural Networks: Graph Classification (Part III)

Read More

Driving Innovation Through Data and Analytics

Read More

Using Machine Learning to Price Apartment Properties

Read More

Turn Your Process Logs Into Actionable Insights With Process Mining

Read More