A few months ago, on a night of infant-induced sleeplessness, I stumbled upon the Reddit group, ‘RepLadies’. This community of fashion savvy women (and men) focusses on finding the best replica designer goods — and we are not talking the ‘Channel’ or ‘Hermis’ littering Bangkok’s night markets and Mumbai’s Linking Road. These replicas are practically exact copies, made of quality materials, and the members go to great lengths to examine how accurate their purchases are. One member even wrote a 30-page guide on how to get the best (fake) Gucci Marmont bags, full of minutiae like the number of button holes on the strap, the placement of the logo and the gap between stitches. These people are not victims, but willing participants in the $460 billion counterfeit industry that is getting bigger, smarter and savvier. The fashion industry though, has found a new ally in its fight against fakes — Artificial Intelligence.
All access
“Cutting edge technology and fashion are not usually seen as kissing cousins,” says Vidyuth Srinivasan, co-founder of Entrupy, an instant authentication service for luxury goods for the resale market. It is essentially a scanner-like device that connects to your smartphone (currently only on iOS). Their patented software is fed millions of microscopic images of authentic and fake bags, after which the device learns the characteristics and identifies them.
The secondary or ‘pre-loved’ marketplace, unlike the original retailers, are far more accessible to the consumer. There are independent sellers with their own websites, eBay or Instagram accounts, which is why brands are anxious to ensure that the goods purporting to be theirs are, in fact, authentic. This information gap between sale and re-sale is the one that another product, Arianee, hopes to fill. The app’s mission is to use blockchain to build the first perpetual, anonymous and trusted record of luxury goods.
- Entrupy
- Pro: Has an immense store of data from brands including Hermes, Celine, Dior, Chanel and more, with a reported accuracy rate of 98.5%.
- Con: Not meant for individual owners but business houses who sell pre-loved goods.
- Arianee
- Pro: Easy to use and provides a unique certificate that can be transferred with a tap.
- Con: The platform has not fully launched yet, so QR codes are not available with warranty cards, and the app is still not publicly available.
- PurseForum
- Pro: PurseForum is the world’s largest designer fashion community with over half a million members. Many of them are veteran shoppers who have the experience and expertise across several luxury brands to instantly determine fake from real.
- Con: The authentication process requires effort (you will have to upload photos, link to your seller, etc) and there have been cases of conflicting opinions.
The technology, which is also used by cryptocurrency, is like a public ledger that is maintained chronologically by several anonymous and trusted accountants. Imagine going on stage and saying, “Hey! I bought this Chanel bag with this serial number!” The audience jots down this information in their ledgers and store it until you want to make a sale. The bag’s ownership will be transferred accordingly. Thankfully, users of Arianee do not need to shout on imaginary stages or understand blockchain technology to use it. Emmanuelle Collet, the co-founder, says, “The QR code will be present in the physical warranty card. Users will just have to scan it and it will automatically be transferred to the Arianee database.” This certificate will be unique and impossible to duplicate, like the IMEI number on your phone. Even if your bag is stolen, you can report that on the blockchain. So if someone wants to legally purchase it, and checks the serial number, they will instantly know that it was stolen.
Dual purpose
While the technological innovations that are heading fashion’s way are increasing by the day, one cannot help but wonder — what happens when good tech falls into bad hands? “AI can make way better fakes,” says Ashwini Asokan, founder and CEO of MadStreetDen, a computer vision based AI startup. “It can actually help learn the patterns and signatures of each brand and recreate them.” Asokan and her team have built one of the world’s first retail-based platforms called Vue.Ai, which helps consumers understand how clothing will look on them, by creating models with their skin tone and body shape. It uses Generative Adversarial Networks (GAN), a new class of AI, to generate these virtual models. Asokan believes that counterfeit goods can be created similarly, although it is going to take a few more years for that to happen.
However, though the design can be replicated, Srinivasan is confident that it will be impossible to create a product of the same quality, given that luxury brands source far superior raw materials, and products are handmade by skilled artisans. The highly-coveted Hermes Birkin bags, for example, are only made by bag makers who have been carefully screened and have trained and graduated from the prestigious Ecole Gregoire-Ferrandi, one of France’s best training institutes.
When it comes to going after counterfeiters, AI is, without doubt, a double-edged sword. But Asokan says, verifying and ensuring authenticity is only scratching the surface in terms of what it can do for the fashion industry. She believes that it is entirely possible that AI can change the way luxury goods are made. “AI brings an insane amount of freedom to consumers to participate in the creation of these goods. And with the growth of 3D printers and generative AI, people can make their own designs and change the face of manufacturing completely,” she says, adding that in the future, the big corporations might not even have the same kind of power that the consumer does.
Counterfeit goods infringe on years of creativity, conceptualisation and craftsmanship. There have also been reports of child labour, trafficking, mafia wars and other related horrors that are now a part of the illegal market’s anatomy. While it is not yet clear how fashion can fully harness the power of machine learning to its favour, AI might just be its most real chance at stopping counterfeits.
