In recent years, the social web has become a vital resource for spreading information on the Internet. News is generated and disseminated on social networks. Products are evaluated and purchases influenced by social review sites. Even mass political action is catalyzed and organized over the social web. However, as the social web has gone mainstream, it has also become an attractive target for organized crime. Viruses and trojans now routinely use social networks to spread, and stolen personal information fuels massive phishing campaigns.
In this talk we will examine three threats to the social web: spam, Sybils (fake accounts), and new phenomena we have termed crowdturfing. These threats are interrelated: crowdturfing systems enable companies to pay small amounts of money to have real human workers generate social spam. These workers, in turn, leverage thousands of Sybils on social networks to spread the spam. We use measured data from Facebook and Renren (the largest social network in China) to demonstrate the scope of these threats. Finally, we present measurement-based techniques to defend against social Sybils, thus crippling attacker’s spam production line. Our Sybil detection technology is actively deployed on Renren, protecting their 220 million users. In the future, we plan to investigate defensive mechanisms that directly combat crowdturfing systems, as well as tackling broader threats to the Web posed by over-personalized “filter bubbles.”
Christo Wilson is a 5th year PhD candidate at the University of California, Santa Barbara, working under Ben Y. Zhao. He earned his MS in Computer Science at UCSB in 2007, and his BS in Computer Science at UCSB in 2006. His research interests include complex networks, online social networks, security and privacy, and networking protocols. He was the winner of the Best Paper Award: Honorable Mention at SIGCOMM 2011, was awarded the Dean’s Fellowship from UCSB in 2010, and received the Distinguished Graduate Research Fellowship from UCSB in 2008.