MIT's Immersion Project Shows What The NSA and PRISM Might Know About You

It’s deeply disconcerting, to say the least, that the NSA has been able to extract data about U.S. citizens from Internet companies using the secret (until recently) PRISM program. Because all of those companies likely handle data a little differently, it’s hard to say how much information the government is able to extract from email records, but a tool called “Immersion”, developed by MIT, shows you what your Gmail account can say about you.

Immersion uses the same data that the government would have access to if it requested it from Google; it’s metadata, which includes the names and email addresses of everyone you communicate with. With that metadata, it’s a simple process to see how much of your email life is allotted to each contact. Immersion uses a helpful bubble-and-line chart to demonstrate it quite clearly.

Immersion chart showing Gmail metadata and what the NSA might now about ut
Immersion chart (Credit: National Journal)

Each bubble, or node, represents a contact. The larger the bubble, the more important that person is to you personally, online. The gray lines represent connections between your contacts. Thus, from a single Gmail account, a number of relationships are illuminated.

The tool wasn’t working for us when we tried it out here--we believe the site is probably just overloaded for the time being--but others have successfully run Immersion and found some intriguing results. For example, Brian Fung of the National Journal found that Immersion dug up 606 inbox collaborators covering about 83,000 emails, and Immersion neatly grouped them all into categories including his college friends, family, friends in a certain geographic area, and colleagues at two different workplaces. The data covers the 5 and a half years he’s had his Gmail account, but he noted that Immersion also lets you break down data by year, month, and week.

There’s a lot of information to be had about you from this very basic information, and it doesn’t even include the content of your emails--which Google already reads, so that data is technically also there for the taking, should the NSA find a way to get it. Imagine what the NSA could learn from crunching this data using multiple email accounts from various connected parties.