- Huw Hughes |
With the increasing integration of AI and technology in our everyday lives, it seems only right that fashion follows suit and adapts itself to modern innovations. One company that seems to be bridging the gap between technology and fashion is London-based retail data analytics company Edited, whose recent study showed that its AI technology was more effective in analysing garments and classifying footwear styles and subcategories than humans.
In a study which put people against machines, Edited’s AI technology made 2.5 fewer errors while identifying garment types, and 9 fewer errors when determining subcategories than their human counterparts. Edited’s data scientist, Kris Graham, told FashionUnited that he was encouraged by the results and their implications for the future of fashion: “Our experiment showed that AI is more accurate, more consistent, quicker and more reliable at scale compared to humans. So even if you had the time to identify the tens of millions of products we see everyday, our algorithms would still make fewer mistakes than humans would.”
AI faster and more efficient at classifying than humans
The results of Edited's study showed that its technology was capable of classifying sample datasets in a matter of seconds, while it took on average six and a half minutes for a human respondent to classify 57 random products - the equivalent of nearly 2.5 hours for the whole dataset.
AI is quick and error-free in classifying large product catalogues, instrumental in analysing competitors' offerings, as shown in the image below.
Earlier this year, Asos - one of Edited’s first customers - attributed a revenue increase of 37 percent in the last quarter of 2013 to Edited’s data insights which helped ASOS to structure its prices competitively.
Graham explained that by embracing AI in retail, the whole industry can be made more efficient: “The results show the level of progress machine learning methods can make to analyse big data. This enables retailers to sell more products with less discounting, understand competitor’s pricing, cut waste, correct product assortments and minimise gaps, along with spot key trends early. It represents a major drive towards efficiency otherwise not possible without the help of AI.”
Edited hopes that its technology - and future AI technology - will be able to do more simple and laborious tasks, allowing people to focus on creative strategies. When asked whether fashion professionals should be worried about machines replacing them in the future, Graham responded that machines still lack the abstract thinking that humans have: “We like to compare our algorithms to dishwashers. It’s a piece of kit that does the boring and otherwise time consuming bits of the job well. It also frees up time so everyone can put energy into doing other creative aspects. But add a dishwasher in your kitchen and it is unlikely to replace the chef. Just like how AI can’t replace humans!”
A future of fashion where AI and humans work together
AI seems to be hot topic in the fashion industry at the moment. On Friday's instalment of TheCurrent Innovators podcast, Ozwald Boateng commented on the rising presence of tech in fashion. When asked about the implications of tech clashing with fashion, the UK designer highlighted the fashion industry’s need to adapt: “The reality is that we have to embrace technology...I think if more designers looked at the world of technology and applied their creativity to the tech world I think we would get some very creative and groundbreaking ideas.”
The UK designer said that a balance would need to be found between the two industries, though the skill of artisans must not be overlooked by tech companies. “I think eventually both parties will realise they need each other and will just make it work. I think a lot of technology companies who are developing this new tech need to have stronger relations with the artisans. They tend to sometimes develop the tech without speaking to the creators who build and the product - there’s a disconnect.”
In a year which saw Italian e-commerce label Yoox Group announcing a new label that is fully developed on the basis of data channelled by AI, American Stitch Fix developing an algorithm for new designs, and IBM Watson working with companies like Tommy Hilfiger and Marchese on the idea of 'cognitive creativity,' it’s undeniable that the reach of technology in the fashion industry is steadily expanding. AI still has a long way to go before the reality of its potential is fully explored, and the fashion industry and technology world need to find a balance - as Boateng said - to reach that point together.
Photo credit: Facebook, Ozwald Boateng/Pexels.