Strata 2012: Is Privacy a Big Data Prison?

I just  returned from O’Reilly Media’s latest soiree, StrataConf 2012: Making Data Work. My panel, If Data Wants to Be Free, is Privacy a Prison? focused on design frameworks for “responsible innovation” to guide data professionals through the perilous territory of data collection, access, and usage. Yeah, I know, ambitious …

I was joined by privacy researcher and NYU doctoral student Solon Barocas [BTW, can’t help but plug Solon’s doctoral chair, Helen Nissenbaum, who had a significant influence on the recent White House Privacy Framework]. Solon tempered my pragmatic, more libertarian perspective with deep thinking—a good combo I think. The session was moderated by the multi-tasker Alex Howard of O’Reilly Media, who was able to keep Solon and I on task and take audience questions from Twitter. Impresssive. You can see the real-time audience discussion on the Twitter #strataprivacy hashtag.

In an effort to steer the privacy conversation away from its predominant (though distracting) themes of data collection and access, fears of a future dystopia, and the complexities of law, our discussion began by examining the implications of big data gaffes through real-world use-cases. Using the Places-Players-Perils Framework lens to focus the cases, we discussed how societal norms should be mapped to the growing power of big data.

And boy, did we have a bunch of recent gaffes, flubs and faux pas to choose from. Take for example, the case of the two cheeky British tourists who were denied entrance into the U.S. for (if only in jest) threats. Or, the FBI cutting back GPS surveillance in the wake of the Supreme Court’s US v Jones decision. Or, that Target knows a teen is pregnant before her parents do. Or, the unification of Google’s services which links all user information across all accounts. Is this 2012 or 1984?

In all seriousness, big data pros can stay off privacy third rails and out of the headlines with a little thought and sensitivity. For example, in the Target teen pregnancy case, Solon made the point that Target could have just asked customers if they wanted coupons during their pregnancy. If the data analysts would have thought about how personal pregnancy is, they simply would have asked. After all, targeted coupons when expecting a kitchen remodel is bit less sensitive than when expecting expecting. Duh.

Even processing of public data can lead to discovery of private information. In fact, from the audience, Daniel Tunkelang asked the provocative question whether the inference of private facts from public data is a new definition of thoughtcrime? BTW, Daniel has a good Strata wrap up today on his blog.

The spontaneous emergence of these weighty issues is why data professionals need shorthand ways to easily think about the societal impact of their work. Such a framework would need to go beyond default “procedural protections” (like the Fair Information Practice Principles) to “substantive protections” that evaluate possible product impact at design-time and track actual impact as the product moves into the market.

More from Jim Adler
Chief Privacy Officer