When Aran Khanna was a college student, he accepted an internship to work at Facebook.

Even before his internship started, he started playing around with Facebook’s APIs and applications. Aran built a Chrome extension called Marauder’s Map, which used Facebook Messenger’s web APIs to track where people lived, what their schedule was, and other highly sensitive information. These were not public features of Messenger, but Aran was able to reverse engineer the APIs.

As a result, of making Marauder’s Map, Aran’s invitation to work at Facebook was retracted. Aran remained curious about the norms of publicly available social network data, and the second order data sets that could be built on top. Out of this curiosity, Aran created a tool called Money Trail, which used public Venmo data to model a graph of how users were paying each other. Aran showed for a second time that data that seems innocent to share can be repurposed to identify, classify, and incriminate users.

Developers of these online applications face tradeoffs between privacy, convenience, and security. By interacting with these applications, we generate data that suggests how we think, what we like to do, and who we are affiliating with. Google and Facebook probably understand you better than you understand yourself.

Aran Khanna previously was on the show to talk about machine learning at the edge. At the time he worked at Amazon Web Services. He now works as a digital privacy researcher. His background in machine learning makes him well-equipped to talk through the subtleties of modern digital privacy. In this show, Aran returns to talk through the finer points of privacy, data, and artificial intelligence.

The post Digital Privacy with Aran Khanna appeared first on Software Engineering Daily.