Mayank Dhiman is now a Powell Fellow in CSE's Security and Cryptography group. But back in 2013, he spent the summer as an undergraduate summer researcher at PayPal, working under CSE alumnus Markus Jakobsson (Ph.D. '97), the company's principal data scientist. Flash forward 18 months, and the two researchers (plus a third former undergrad intern, Hossein Siadati) had their worrk on Liar Buyer Fraud published and accepted for presentation at the Workshop on Usable Security (USEC 2015), which took place in early February.
In the paper, Dhiman and his colleagues describe a common yet poorly known type of fraud -- so-called Liar Buyer fraud -- and they go on to explain why traditional anti-fraud technology has failed to curb this type of fraud, which typically involves a purchase gone bad. Typically the consumer places an order and receives the merchandise, then reports it as not delivered, and asks for a refund. The researchers introduced a counter-intuitive technique in which the claims processor unveils knowledge of detailed iniformation about the purchase, and new experiments showed the new technique based on user-interface modification to address liar buyers has the potentialto dramatically reduce losses from fraud. Using a combination of role playing and questionnaires, they determined the opinions and behaviors of about 1,700 subjects, and found that their proposed technique results in a statistically-significant reduction in fraud rates in an experimental setting. The results are so promising for both men and women that the researchers hope to expand the study to real e-commerce traffic, when the researchers could also test whether men are more willing to lie and defraud than are women. While research has previously confirmed that discrepancy, Dhiman and his colleagues say that results using their new counter-intuitive technique so far show men being about as honest as women.
Read the paper on Liar Buyer Fraud, and How to Curb It