August 25, 2014
In a step toward protecting the personal data of online users, researchers at Columbia University have created new software called XRay that can observe and predict how tech companies are using the personal data that they collect. The software is based on research related to Google’s Gmail ads, Amazon recommendations, and YouTube recommendations. XRay, which will help privacy-concerned watchdogs track how personal data is used, is still in development.
The researchers studied how different inputs of data, such as emails, searches, or product views, affects the outputs like ads or recommendations. For example, they found that email messages about pregnancy brought up ads for baby shower invitations and Ralph Lauren apparel. In messages related to depression, they found a correlation to ads for shamanic healing.
By leaking the targeting methods, the researcher hope to bring more transparency to a process that uses massive amounts of personal data from consumers.
“That leaked targeting information can potentially be used for all sorts of purposes. It can be used for discrimination. And it is a very hidden kind of discrimination,” Roxana Geambasu, assistant professor of computer science at Columbia University, said in The New York Times.
The prototype of XRay will be shown at the Usenix Security Symposium this week, but the open-source tool won’t be released for at least another year. Early adopters may include privacy groups, state attorneys general offices, the Federal Trade Commission, and even journalists.
Researchers at Columbia received grants from DARPA, the Brown Institute for Media Innovation, the National Science Foundation, Google, and Microsoft to continue work on the XRay project.
For more information, visit the XRay project site.