This research, which was begun in 2013, studied the Internet privacy of adults and children. The focus of the research was on how 3rd parties -- such as foreign governments and companies such as the now infamous Cambridge Analytica --  can scrape information from social networks, search engines, and public records, combine the information, and use data science inference techniques to obtain very vivid pictures of the private lifes of large swaths of adults and children. Our research group was among the first to recognize this problem and undertake an indepth inverstigation of potential 3rd-party attacks. Our research showed how two privacy laws -- COPPA and the Right to Be Forgotten -- originally intended to improve the privacy of children and adults could backfire and acutally reduce Interent users' privacy. We showed how parents who post photos of their babies and children can inadvertently leak - through inference techniques - private informaiotn of their children. We showed that anonymous usesrs on Twitter exhibit significantly different behavior than non-anonymous users. We showed that location-based social networks can often be exploited by 3rd parties to determine the exact location of users, and to de-anonymize users in anonymous services like Yik Yak. The research led to 13 research papers on privacy and anonymity, covering a variety of different Internet applications, including Facebook, Google, Twitter, eBay, mobile dating services such as Tinder, and anonymous mobile apps such as Yik Yak. Our work has been extensively covered by the media, including the New York Times, NPR, and Bloomberg.

We are not at all surprised that many Internet applications today are being successfully attacked by thrid parties, such as foreign governments and companies working for political campaigns such as Cambridge Analytica. In particular, because Facebook over the years has been overly aggressive in trying to network the world, it has set itself up for numerous privacy attacks. Although Facebook has recently taken small measures to reduce the problem, it needs to make more fundamental changes to provide the safety and privacy that Internet users expect. In particular, we recommend that Facebook follow the example of Wechat, the popular social networking and messaging applicaiton in China. In Wechat, users cannot see other users' friend lists, and it does not recommend potential friends to users. By making this fundamental change, Facebook would go a long way in reducing potential inference attacks that currently plague the social network. 

We would like to thank NSF for its generous support to pursue this research. 

Last Modified: 10/16/2018
Modified by: Keith W Ross