Experiences that test the power of Artificial Intelligence to curb counterfeiting are multiplying. From the case of the Patou brand to the second hand, we take stock of how AI can become essential in uncovering fakes and making not only their dissemination but also their production more complicated
At Case Western Reserve University in Cleveland, USA, they traced the author of a painting by analysing a tiny portion of the work. Even the diameter of a brush bristle was a crucial element in reliably identifying artists. This is the power of Artificial Intelligence. Art is just one of the fields in which it can be applied. But it is clear that the same technology could work with a leather bag. Either at the production stage to protect it from being copied. Or to expose fakes already in circulation on the market. Luxury brands and second-hand companies, not by chance, are investing precisely in this direction to solve a problem that until now seemed unsolvable: counterfeiting.
The Power of Artificial Intelligence
Patou, a brand in the LVMH orbit, is protecting the limited edition Le Petit Patou bag model so it will not be copied. It is the first test to extend the system to other brand products. LVMH is at the window to observe the results and assess whether to apply them to other brands in the group. What system does Patou use? It is called Authentique and was created by Ordre, an online wholesale platform. During the production stages of the bag, a photograph is taken of a specific area of the product to create a fingerprint. Authentique’s algorithm analyses the image at a microscopic level, translating the data into a numerical code and creating a digital ID stored as an NFT so that it cannot be altered.
The customer who buys this bag enters the product code into the Authentique app, which will guide them to take a picture of the product at the required location and compare the code generated from the image with the one on file, verifying its authenticity. One of the advantages is that there is no need to apply unique tags such as NFC chips, which can be removed or counterfeited themselves. This system now applies to clothing, shoes, and small leather goods. In a year, watches and jewellery will be added. The accuracy rate in distinguishing real from fake is over 99% and, the developers say, is improving.
The greatest challenge
But the biggest and most demanding challenge for Artificial Intelligence lies in detecting fakes already on the market. One example: Entrupy was founded in 2012 with this purpose. It developed its AI through a huge archive of data collected from authentic and counterfeit items dating back decades, including riding trunks and whips from the 1930s, produced by companies that later became luxury giants. Entrupy provides an optical device that allows a smartphone to function like a microscope.
Photos are transformed into data compared with those in the archive to unmask any fakes. On the front of AI as a weapon to curb the scourge of counterfeiting, the big players of the second hand have already taken action. According to Business of Fashion, StockX, and The RealReal are both already using it, implementing systems that learn and improve thanks to continuous data input, making them increasingly reliable.
Chapter blockchain
Another chapter closely related to the previous ones: the interconnection between AI and blockchain, which has enormous potential and is attracting the interest of several sectors. One is that of cybersecurity. Artificial intelligence can be used to detect and respond to threats, while blockchain technology can ensure the security and integrity of data. Another application area is that of the supply chain. Blockchain can create a transparent and secure one, while AI analyses data and optimises the supply chain. This enables companies to reduce costs, improve efficiency and ensure that products are delivered on time.
The human factor
In all this, what is the role of man? At Case Western Reserve University, they say that the human factor is often overlooked in this area. Mistake: the system does not work alone, and you need people with extensive product knowledge to collaborate and improve the artificial interface. On their own, machines are useful, but they are of little use.
Read also: