The Impact of Data Integration on Privacy

Source: ACLU. Voices: Jason Stiles and Melissa Schwartz

I like to include the ACLU's 2003 video about "ordering pizza in the future" in most of my presentations about privacy. From the first time I saw it, there were two points that stuck out to me.

The first being the power of integrating silos of information to gain insight in customer needs and preferences. In a positive case, you can image a retailer knowing that you will be going on vacation in a week and that you have bought designer clothing in the past being able to highlight deals and discounts on fine clothing for tropical climates. However, as the video points out, every technology (or intention) has positive and negative sides. The line between social good or convenience ("positive") and creepy
 ("negative") is not well-defined, but is intuitively known by the consumer.

The second point is the decreasing levels of privacy inherent when one integrates more and more data (even if one is "only" merging metadata). In the process of collecting more information about someone, it becomes more likely than one can identify that particular person in the large crowd. The collected attributes that describe them and their behavior form a unique fingerprint, which is evident even when portions of the data are de-identified.

It is the second point that concerns me most. The current push is towards big data analytics and using cloud technologies whenever possible. This focus puts a spotlight on the erosion of privacy.

Unfortunately, there is more emphasis in the mainstream discussion on the potential benefits of this information explosion and very little practical technical or policy action on guarding against the possible negative outcomes.

To quote Bruce Schneier, "Data is the Pollution of the Information Age". Shall we wait, like the previous generation, until it is too late to address the problem?