How AI can spot counterfeit luxury goods
When shopping for your favorite brands, how easily do you think you could spot a fake? Sure, if you see a nice watch or handbag on the street going for a fraction of what it’s worth, it’s pretty easy to tell it’s a knockoff. But if you had to compare two products, one a counterfeit and one the real thing, could you actually tell the difference? It’s difficult to do. That’s why counterfeit goods account for nearly $600 billion in sales every year, making up 3.3% of global trade. AI companies are setting computer vision to the task of sussing out the fakes.
Scan Verification
Thanks to machine learning, AI programs can train scanners to pick up on the subtleties that mark the genuine article. Things such as number of stitches on seams that indicate a high-quality luxury item as opposed to a cheap knockoff that might fall apart in a few years. Not only do counterfeits dilute a brand, they’re of poor quality and not nearly as durable as the real thing. Luxury goods aren’t just a fancy design or label, they made their reputations on first and foremost making quality products.
Entrupy claims a 99.1% success rate in authenticating luxury items, giving wholesalers and buyers peace of mind on a purchase. The Entrupy app scans goods from multiple angles for the telltale signs of the real thing. Its algorithms, armed with more than 15 million unique images from authentic and counterfeit goods, assess the scans and provide a determination in 60 seconds or less.
“To remove human error from the equation and accommodate a wide range of businesses by automating authentication through machine learning and AI. The company spent its first four years just collecting data from all over the world, which largely involved buying a lot of luxury designer handbags — both real and fake. We spent a substantial amount of money on the best counterfeits and then also buying from the brands,” then-Entrupy director of business development Deanna Thompson told Future Stores.
Ennoventure helps brands prove they’re the real deal by installing encrypted signatures on packaging. Purchasers can scan the packaging to make sure what they bought is what it claims to be.
“Basically, it’s a software solution which will help you to check the validity of product packaging. Its encryption technology combines AI, cryptography, and blockchain to create an invisible encrypted code that counterfeiters won’t be able to read,” CEO Padmukamar Nair told SME Futures. “On several fronts, multiple machine learning models are used to detect questionable signals. A deep neural network, for example, can swiftly detect fake product lookalikes, while blockchain stores the product’s e-pedigree from the manufacturer to the consumer.”
Unlike QR codes, the signatures can’t be copied, and unlike hologram stickers, they can’t fall off the packaging.
Cleaning Up Shop
Most counterfeits these days aren’t sold on the streets, but online. The pandemic-driven e-commerce boom has served to exacerbate the problem of fraud. And unlike street vendors who are selling knockoffs at low prices, online counterfeiters are advertising the real item at the real price only to ship a lesser-quality fake.
This obviously hurts buyers who got scammed, but it also harms legitimate brands whose products and images are used to bait those buyers before the switch comes. That puts pressure on online retailers to guarantee some authenticity.
In response, Amazon launched Brand Registry to enable sellers to manage their product listings and make sure only legitimate items are listed as theirs. The registry also proactively removes suspected knockoffs. Its Project Zero AI tools scan Amazon listings and remove suspected counterfeits. Brands can also remove any counterfeit listings they see themselves.
Chinese e-commerce giant Alibaba has even more incentive to chase down and remove counterfeits given that China’s government holds online vendors liable for any fraudulent products sold on their sites.
“Over recent years Alibaba has really stepped up protection and really knuckled down against bad actor accounts on the site and infringing listings, especially during the pandemic,” Claire Jones of intellectual property law firm HGF said at September’s Luxury Law Summit in London. “A lot of this stems from the increasing pressure from the Chinese government and various regulatory crackdowns they are undertaking to improve the image of China and counterfeits.”
Fraudsters continue to get better at mimicking genuine products and their packaging, and are harder to detect as they find more outlets to hawk their counterfeit wares. Computer vision and machine learning technology must keep improving to spot increasingly sophisticated methods. As always, the buyer should beware, but now those buyers have more tools to sniff out fraud.
Leave a Reply