By creating a unique “fingerprint” of an item, AI-based software could offer a solution to fighting counterfeit goods, which cost the world’s economy hundreds of billions of dollars each year.
Hologram stickers, QR codes, and NFC chips are just some of the tools companies use to prevent counterfeits. They do not work – and the global counterfeit problem is only getting worse, according to Roei Ganzarski, head of Alitheon, a company that seeks to reverse the trend.
“The technology that the counterfeiters use is fantastic. It gets better and better every year as well,” Ganzarski said, adding: “it’s just like the virus and anti-virus problem in computers.”
OECD, a club of mostly wealthy nations, said counterfeiting was a “worrying threat” that accounted for 2.5% of the world’s trade, or $464 billion in 2019, according to the latest available data. It is equivalent to the entire national output of countries like Austria or Belgium. And that’s a conservative estimate based only on customs seizure observations.
The International Chamber of Commerce projected the counterfeit and pirated goods market to reach a whopping $2.8 trillion this year. According to Forbes, counterfeiting is the largest criminal enterprise in the world, more lucrative than illegal drug trade or human trafficking.
Based in Bellevue, Washington, Alitheon was founded in 2017 to create a tool that could help tackle fraudulent production. It has developed an optical AI software called FeaturePrint, which it says can spot fakes with 99.85% accuracy and zero false positives.
The program takes advantage of what is called “tolerance” in manufacturing and industrial engineering – or permissible variations in the final product that fall within the margin of error.
It means that machines cannot do the exact same thing over and over again, Ganzarski said, so no item ever made is the same as the one produced before or after it, even if they seem identical. Which means each has a unique “fingerprint.”
“Our algorithm looks at the item and finds small features, small flaws of the manufacturing process that are inherently and naturally there and are random and chaotic by nature. We find them. We create a digital fingerprint out of them,” Ganzarski said.
A standard phone camera can be used to take a picture to authenticate the product. It makes it an easy-to-use and cost-effective solution, according to Marcus Behrendt, managing director at BMW i Ventures, which has recently become Alitheon’s principal investor after announcing a $10 million financing for the company.
“As of today, we help ourselves with serial numbers and barcodes or, in more important cases, with chemical trackers to truly identify an object. Alitheon can do that just with a picture of the surface of the object in question,” Behrendt said in a statement to Cybernews.
Beyond money matters
BMW i Ventures co-led Alitheon’s latest investment round with Seattle-based company Imagine Ventures. Operating from Silicon Valley, a German car giant’s venture capital firm also made a small investment several years ago to help develop the fake-spotting AI software.
“It is a very versatile and powerful tool and can basically be applied anywhere. But especially at valuable or safety-relevant applications, this technology will demonstrate its true capabilities,” Behrendt noted.
While fashion and luxury is one obvious industry that could benefit from a technology like FeaturePrint, it could also do more than save billions of dollars companies spend on fighting fakes. It could also be a life-saving solution in sectors like automaking, the aerospace and defense industry, or medicine.
“Imagine car brake pads, for example, where the brake pad is fake, but the box or the hologram sticker on the box looks real. Then you can find that you, by mistake, put a fake brake pad into a car, and then an accident happens – people can get hurt or even die,” Ganzarski said.
Aside from spotting fakes, FeaturePrint, a patented technology, can also minimize the odds of part misuse and trace the provenance of high-value products like gold bars. While there are chemical ways to check if the gold itself is real or not, the issue of origin remains.
“The big challenge today is if the gold is legal. Is it coming from Russia, is it coming from the Congo, from Uganda – where is the gold coming from? Is it from places where the mines do a good job with employees? Or is there child labor, illegal things going on?” Ganzarski said.
More from Cybernews:
Subscribe to our newsletter