UMD Researchers Offer Insight on Social Media Privacy

Your online activity, whether tweeting, photo sharing on Facebook, or playing Candy Crush or FarmVille, often provides an intimate portrait to friends and followers of who you are and what you like to do.

These same digital brushstrokes, however, can reveal much more information than intended, allowing companies and organizations to learn everything from your personality traits, to your alcohol use, to your sexual orientation.

A pair of University of Maryland researchers will offer insight into social media privacy—as well as data mining techniques used to track online users’ behavior—in a panel discussion this Saturday at the eclectic South by Southwest (SXSW) conference and festival in Austin, Texas.

Philip Resnik, a computational linguist in the University of Maryland Institute for Advanced Computer Studies, and Jennifer Golbeck, director of the university’s Human-Computer Interaction Lab, will be joined by two other experts to examine questions like: why, how and when do our digital lives expose us to social media experiments and invasive snooping; how do these activities affect users in real life; and what ethical issues arise in social media experimentation?

The panel, organized by Resnik, also includes Jason Baldridge, co-founder and chief scientist at People Pattern, and Michelle Zhou, co-founder and CEO of Juji Inc.

“I’m aiming for a discussion with a lot of energy, that hopefully will leave the audience better informed about the technology, the kind of research people are doing with it, the value of that research, and the risks and challenges associated with it,” says Resnik, a data mining expert whose recent work includes how the context of spoken words can help identify depression in young adults, and developing big data analytic tools to track different political opinions on mobile devices.

Golbeck, a computer scientist in the university’s iSchool, gave a popular TEDx talk in 2013 called “The Curly Fry Conundrum” that detailed how scientists are building algorithms that can infer almost everything about you, just by examining your “likes” on Facebook.

“Employers, insurance companies, the police—they could all potentially use this technology in ways people might not like,” Golbeck says.

In light of her own research—and the backlash to Facebook’s 2012 attempt to study “emotional contagion” by manipulating users’ newsfeeds—Golbeck says she posts only a minimal amount of information on her own social media accounts.

“Other than the occasional thing that I very consciously think, ‘I am going to keep this on my profile,’ I do delete everything after two weeks,” she says. For example, she's kept her wedding photos up, but her tweets about Washington Capitals hockey games come down after they're no longer timely.

Read more about the panel discussion here.