FYI.

This story is over 5 years old.

Tech

Snoop On Your Own Metadata With MIT's Help

MIT Media Lab's César Hidalgo, Deepak Jagdish and Daniel Smilkov are allowing email users the chance to look at and learn from their own metadata.
A screenshot of the Immersion website visualizing user metadata.

Metadata. A month ago, the average person wouldn't have given a microsecond's thought to the word. Hell, most Americans probably didn't even know the word existed, let alone the data. But, in the wake of Edward Snowden's NSA leak, everyone has an opinion on metadata's privacy implications, as well as exactly who has access to these rich, global etheral mines of time, data and location information.

For the US government and President Obama, in particular, it has become political dynamite. On July 4, for instance, a group called Restore the Fourth is planning nationwide rallies, while European allies are collectively arguing that the snooping will strain upcoming US-EU trade talks. "Your data is gold," Thomas Drake said at a MoMA event on Sunday.

Advertisement

Now, a trio of geeks at the MIT Media Lab are attempting to give users a chance to actually embrace the big metadata gold rush and learn from themselves. In the process, César Hidalgo, Deepak Jagdish, and Daniel Smilkov hope to depolitize data mining and better demonstrate its value -- and costs -- to the end user.

Hidalgo, a profesor at the MIT Media Lab, where he is head of the Macro Connections group, is primarily interested in human understanding of and "developing concepts of complexity, evolution and network science." Jagdish is the large-scale data guy, while Smilkov previously worked in the field of network science. The three bill Immersion as "an invitation to dive into the history of your email life in a platform that offers you the safety of knowing that you can always delete your data."

The group sees the project in almost artistic terms. They see the new internet user experience as a blank "canvas" that acquires data over time. Comparing Immersion to a "cubist painting," they consider the site a form of "artistic representation that exists only in the presence of the visitor." When this data is arranged visually, like the tentacles of an octopus, it's easy to see why. The idea, to quote from an Immersion video, is to "allow you to see the web you have weaved together with others."

Jagdish notes that Immersion orginally came into being as an interactive installation at the Media Lab. It featured a guestbook graph and a ranking field that is currently unavailable on the website. When visitors logged into the installation's guestbook, they appeared as a node onscreen. The ranking view, on the other hand, visualized the users' most frequent collaborators (those people to whom users sent three or more messages.)

Advertisement

Immersion: Beneath the surface from Deepak Jagdish on Vimeo.

Jagdish, Hidalgo and Smiklkov also see "Immersion" as an opportunity to more strategically and efficiently interact with professional contacts, which is an interesting concept. Whether this can help users weather the deluge of personal communications from family and friends is yet to be determined. We all could use that sort of help.

Of course the premise of "Immersion" is as old as the social network: Google, Facebook, Yahoo and an untold number of entities are constantly monitoring the metadata that represents our online behavior, in ways that we often barely understand. Buried inside all of the big Internet's terms of service is text that allows those companies to yield your data -- and not just your metadata -- to the government. The NSA's wiretapping system seems to have more direct access to those servers. But only the technology is new. Using your data to better understand you is an old practice of advertisers and the government; you just haven't always reaped its benefits yourself.

The catch is that "Immersion" users will have to share their Gmail logins and passwords with the site. The creators of "Immersion" guarantee that it is a secure system, just as Facebook, Google, and the other miners of our metadata (and other kinds of data) claim. The difference here is that users can return to the site at any time and, according to its creators, delete their shared metadata. A novel concept, for in the real world, metadata is permanent—there is no erasure. But, it's an interesting thought experiment: an internet in which we could delete our own data in a complete way, preventing it from being served up to the prying eyes of corporations and government spy agencies. That's a nice thought.

The project has yet another benefit that Hidalgo and his compatriots fail to emphasize, perhaps for obvious reasons. Users curious about the true nature of metadata will see that "Immersion" cannot access actual email content. (A project launched by Miranda July this week promises to publish random emails from celebrities' inboxes each week, starting with Lena Dunham.) "Immersion" can only visualize what the metadata actually delivers—date, time and location of email communications. For that reason, it could be a good educational tool in understanding the ways in which the NSA and other agencies may be making use of its metadata: that is, looking for patterns, building surreptitious social networks. If the Media Lab pitched "Immersion" in this way, though, they might be chased off the internet.

Still, even with just our little bits of metadata, "Immersion" offers an interesting opportunity to snoop on one's self. And it's perfectly constitutional.

To sign up for the project, head over to immersion.media.mit.edu