When Big Data Breaks Bad

When Big Data Breaks Bad

MIT's Engaging Data Conference fails to engage on pressing issues

  • Posted on: 29 November 2013
  • By: stannenb

MIT's Senseable City Lab held a wide ranging discussion in November seeking "a new way to frame the relationship between individuals and Big Data, to move beyond today's pseudo-feudal system of trading personal information for a service." While there was interesting conversation at "Engaging Data 2013: Big Data or Bad Data" this gathering of "thought leaders" missed an opportunity to engage on the most difficult of issues, largely by ignoring MIT's own role.

The day began with a conversation between MIT Institute Professor & Professor of Linguistics (Emeritus) Noam Chomsky and Washington Post journalist Barton Gellman. Chomsky, who helped Daniel Ellsberg leak the Pentagon Papers, and Gellman, a recipient of Edward Snowden's leak of NSA documents, replayed the sort of conversation that Chomsky, a radical critic of American foreign policy, has had with liberals for decades. Citing historical examples such as the American counterinsurgency in the Philippines a century ago, Chomsky said that the NSA surveillance operations were another example of how governments always used the most advanced technologies to surveil and repress. Chomsky noted that, while existence of the programs didn't surprise him, their vast scale did. Gellman said that one didn't have to assume "bad faith" to criticize the NSA, and that the officials who set up and operate these programs were, in good faith, seeking to protect America from what they felt were genuine threats. Chomsky then listed an array of "monstrously repressive" governments whose officials said the same thing. Noting that America was far more free than any of those regimes, he said that "in a technical sense, those statements [from American officials] have no information content" and deserve to be ignored. Gellman, seemingly trying to find some common ground between himself and his fellow critic of the NSA, called for more transparency. Chomsky, who has written books about corporate control of media and state propaganda, demurred, saying that transparency was insufficient in an environment where critical stories go unreported. In response to an audience question about whether we needed these surveillance programs to protect against terrorism, Chomsky said that, if the US government genuinely cared about terrorism, it wouldn't be engaged in what he termed the largest terrorist campaign ever, drone strikes against suspected terrorists.

It wasn't clear, however, what light this discussion shed on "Big Data". While it's certainly true that the NSA's collection of data is massive and that they're relying on many of the same tools as Google and Facebook, Chomsky's point that governments tend to collect dossiers on citizens begs the question on how these collection efforts are different from what's come before.

Saskia Sassen, Robert S. Lynd Professor of Sociology at Columbia University and the day's sole female speaker or discussion leader, followed with a set of points which, if connected, would have made a far more powerful provocation. Sassen spoke of the "materiality" of the surveillance state. Citing the Pulitzer Prize winning work of The Washington Post's Dana Priest and William Arkin, Sassen talked of the physical installations that support the monitoring of America and asked the audience how many were aware of its vast scale. Just a few hands were raised. Sassen went on to make a separate point about how marginalized populations have knowledge than can and should be harnessed, viewing that as an empowering act. But it is the marginalized and vulnerable populations who are quite aware of the growth of surveillance and control. Immigrant and undocumented communities have felt the scrutiny the Department of Homeland Security's Immigration Services unit. African American communities, beyond the stop and frisk policies in New York, have felt the disproportionate impact of the war on drugs and, with the growth of criminal record background checks, have had their employability hampered long after criminal sentences were served. Sassen's joke that the NSA's tapping of Angela Merkle's cellphone was going too far would have been funny if she had heeded her own methodological advice, listened to the wisdom of the marginalized, and realized that surveillance had gone too far long before it was listening in to the conversations of a head of state.

The rest of the day was spent in small discussion groups, seeking to explore further some of the topics raised during the day. What wasn't discussed, however, was in some ways more interesting than what was. At last years Harvard Law School Symposium "Privacy & Technology", an array of lawyers, complaining that law was lagging technology, advanced a far more knowledgeable technological critique. Rather than worry about whether we should "trust" Facebook and Google, a topic Friday, the HLS symposium considered the growing capability to link online and offline data. Beyond the NSA, the Harvard symposium discussed intelligence "Fusion Centers" and whether these centers, which bring multiple police agencies together with private corporations under one roof, have irrevocably blurred the lines between state and corporate power. While the lawyers at the Harvard symposium might have thought that the facts had outpaced the law, they demonstrated a far more current understanding of those facts than many at the MIT discussions.

From Sassen's perspective of materialities, the unsaid overshadowed the said. We were in the shadow of Kendall Square, where a state initiative and tax subsidies are helping to build a "big data" industry. MIT's Computer Science and Artificial Intelligence Lab has its own corporate sponsored "big data" initiative. The biggest materiality of all, though, is MIT's Lincoln Lab, where MIT conducts research funded by the Departments of Defense and Homeland Security where, among other things, "novel analytics for handling high-dimensional datasets, specifically graph analytics and techniques for fusing and analyzing data from multiple data sources" are developed. Or, to adopt this symposium's formulation, it's where MIT develops technologies that can turn big data bad.

It is, perhaps, too much to expect MIT to engage in such critical self-examination, even in a symposium headlined by Noam Chomsky. But other opportunities were lost. Gellman talked about discovering a mysterious box on his New York apartment building. Investigating, Gellman learned that this was a monitoring device, placed by the City of New York, to track energy use down to the apartment level. Meant to be part of New York's response to climate change, Gellman noted that it would also be of interest to law enforcement should they seek to establish someone's "pattern of life", when they wake, go to work, return home, and go to sleep. That sensor box, and the infrastructure behind it, something of which we will increasingly see as cities react to the climate crisis, is an inflection point where (good) Big Data can become (big) Bad Data. A discussion of that tension, of how a senseable city remans sensible, using the promise of big data while protecting the rights of its citizens, would have been a welcome beacon of light in this otherwise murky gathering.

Videos of the symposium are available on its web site.

Read CambridgeHappenings, a daily Cambridge news summary, curated from fresh, local sources.