Brand Protection: Combating Online Counterfeit with AI
- Brandon Li, Product manager at Austreme
- 18.10.2018 09:15 am undisclosed
Besides the huge financial losses caused to the industry, counterfeit selling is also found to be an important source of income for terrorists. According to Edward Fox, the Customs and Border Protection's acting port director of the Port of New York and New Jersey in Newark, “those dollars (income generated from counterfeit selling) are going to support drug trafficking organizations, organized crime and in some cases terrorist organizations”. For example, “terrorists who attacked the French magazine Charlie Hebdo in 2015 financed their weapons partly by selling fake Nike sneakers”. On a personal level, buying counterfeit goods such as fake drugs is illegal and potentially harmful to health, for example, resembling Viagra is a recently rising fake drug. Thus, brand owners, authorities and payment processing organizations are very serious in combatting IP infringement activities.
To discover and monitor the brand-damaging content and counterfeit on the internet, web crawling is a technology commonly used by brand owners as well as banks and payment companies, which have to fulfill the requirements of card schemes like Visa GBPP (Global Brand Protection Program) and Mastercard BRAM (Business Risk Assessment and Mitigation Program). During the merchant onboarding and monitoring process, many banks use web crawling tools or employ third party service providers to scan the merchant site to see if there is any high-risk content (e.g. pornography, gambling, drugs, counterfeit etc), security loopholes and other risk concerns. Comparing with manual eye-ball checking, web crawling can be more efficient and cost-effective.
Nevertheless, while web crawling technology helps to save a lot of time and human resources in checking an e-commerce website, a weakness of most of the current web crawling tools is that they can only do text-based crawling. That means bad merchants can easily bypass the keyword monitoring by using some ambiguous description with product photos. For example, we see many counterfeit sellers use “luxury”, “top quality”, “Europe brand”, “France brand”, “famous brand” etc to describe the IP infringing products. Buyers can easily recognize the branded item with the product pictures but a web crawler does not. Therefore, many illegal and counterfeit items were missed if brand owners and banks rely on traditional text-based web crawling methods.