Like many others in Singapore, you are no stranger to shopping online. It’s simple, cheaper, fuss-free and delivered straight to your doorstep. In times of COVID-19, thank goodness for online shopping! But how can you be sure that what you have just bought is a legitimate product, containing permitted substances within approved levels?
The Problem
COVID-19 has brought about a significant rise in the number of online transactions taking place on e-commerce platforms, including that of contraband items. As e-commerce gains traction, quickly becoming a way of life for many, some consumers may be caught unaware that what they are purchasing could be illegal. Even if consumers do scrutinize the list of ingredients of a product before making their purchase, they may not know the scientific name of certain ingredients, or they may unwittingly purchase a counterfeit product that contains substances banned locally, designed with similar packaging to the original. Some ingredients, while allowed in small quantities, may be illegal once present in quantities exceeding a certain amount, and it is exceedingly difficult for consumers to keep track of all the various regulations in place for different chemicals. Consumption or application of such items could be risky to buyers’ well-being, resulting in undesirable outcomes.
To safeguard the interest of consumers, officers from enforcement agencies traditionally have to manually trawl through several e-commerce platforms to identify such products and contrabands, analyzing information of these items one at a time. For example, between January and May 2021, the Health Science Authority (HSA) removed more than 3,200 listings of illicit health products on local e-commerce platforms in a crackdown, which was time-consuming and repetitive.
The Solution
Enter the E-Commerce Surveillance Tool. Designed by teams from HTX’s Data Science & AI (DSAI) and Q Team Centres of Expertise, the tool is an AI bot that enables enforcement offices to sift through online listings of contrabands using Robotic Process Automation (RPA) and AI technology.
Wei Lin Tan, Q Team Engineer, testing out the E-Commerce Surveillance Tool (Photo: HTX)
At the click of a button, the bot fully automates the data analysis process, increasing the efficiency and speed of the data mining processes. The tool successfully overcomes unique challenges that enforcement agencies currently face when monitoring for the sale of contraband items, such as having processes that are not streamlined and require different software, ensuring the relevance of listings from search and obtaining text information. This enables enforcement officers to sift through the contraband listings within hours, compared to what would typically take days or even weeks for them to complete.
Future Possibilities
The tool opens possibilities to monitor the sale of a wide range of contrabands – from drugs to illegal or counterfeit health and cosmetic products, controlled and prohibited weapons, and even unauthorised COVID-19 test kits and vaccines. HTX is working with HSA to customise this tool for HSA's enforcement officers to identify illicit health products listings more accurately, and in so doing, take down such listings faster and prevent consumers from unintentionally purchasing and consuming these harmful products. The tool will be trialed by HSA starting 8 July 2021 for a month, followed by feedback and iteration before it gets deployed.
Beyond detecting illegal health product listings, the tool could be deployed within the Home Team to help solve crimes and enhance public safety and security. Several Home Team agencies are interested in customising the tool for their own purpose and will be trialling it in the coming months.
Jason Ng, DSAI Data Scientist, at work developing the E-Commerce Surveillance Tool (Photo: HTX)
The project’s success was the result of close collaboration between members of two HTX CoEs, DSAI and Q Team. Wei Lin, an engineer from HTX’s Q Team CoE, contributed towards customizing the AI technologies behind the tool. She improved the tool’s text and image analytics processes, helping the bot to filter and identify relevant listings accurately and efficiently. Jason Ng, a Data Scientist from HTX’s DSAI CoE, was another member of the team who played a significant role in the formation of this tool and is excited to see it put to good use. Says Jason, “Personally, I definitely felt great that we managed to help to automate the analysis of the data to find contraband items on e-commerce platforms. While working on this project, we realized that contraband items are actually more easily available for unwitting members of the public to buy online than one might expect. Knowing that our work on this tool can help enforcement officers more quickly identify contraband products and thus keep the public safe is very satisfying.”
While Jason had initially been trained in another field, having pursued a major in Engineering Systems and Design and a minor in Artificial Intelligence in Singapore University of Technology and Design (SUTD), he picked up Robotic Process Automation independently from his studies due to interest. In October 2020, seeking to apply his technical skills, he joined HTX as part of the S&T Associate Programme. He found working on the e-commerce surveillance tool project a good way to apply his technical knowledge. “I learnt Robotic Process Automation on my own, as part of a prior internship. It’s quite a niche area, and so I felt very fulfilled when I was able to hone my skills further by applying it to real-world scenarios that directly impact the public.”
While e-commerce sites may appear relatively straightforward, a lot of work goes behind keeping them free from contraband items. So when you engage in your next online shopping spree, know that data scientists and engineers like Jason and Wei Lin have worked tirelessly behind the scenes to make your experience significantly safer!
The people involved in this project are:
Q Team CoE:
Terence Tan - Deputy Director, Q Team
Tan Wei Lin - Engineer, Q Team
DSAI CoE:
Sylvia Liaw - Head of Cognitive Analytics and Ops Capabilities, DSAI
Jason Ng - Data Scientist, DSAI
Kiat Kai Lee - Engineer, DSAI
The E-Commerce Surveillance Tool was also covered in the Straits Times on July 29, 2021.
HSA trials bot that targets contraband items sold online
https://str.sg/3UiN