Transparency in Data Mining: From Theory to Practice - Discrimination and Privacy in the Information Society - page 290

autonomy - a notion that itself carries various meanings in this context. Within
this discussion, the chapter explains how the relevance and strength of these
theories varies in accordance to the segments of the process which were
previously drawn out. Part 5 concludes by bringing the findings of the two
previous sections together. It explains at which contexts the arguments for
transparency are strongest, and draws out the implications of these conclusions.
Finally, Part 6 sets forth a brief coda, which acknowledges that transparency still
leaves many important issues unanswered.
17.2 Predictions, Data Mining, Personal Information and
Information Flows
Governmental predictions call for the use of sophisticated computer programs and
extensive datasets, as well as a role for professional experts and data analysts. The
process relies upon the success of specific technological processes. It is also
premised upon assumptions; some statistical and some pertaining to society and
human nature. To understand these points, the following analysis begins by quickly
examining a famous example of governmental predictive tasks. Thereafter, the
chapter examines the technology enabling these projects, the role of the human
analyst in what seems to be an automated process and the policy decisions
underlying many of the steps of these processes. Understanding the intricacies of
these processes is crucial for establishing the importance of transparency, and how it
could be applied in practice.
17.2.1 Example: Data Mining and Security 2
Since 9/11 and subsequent attacks around the world, governments are working
extremely hard to preempt such events. Among various initiatives, it is reported that
governments are employing predictive data mining to study trends in the actions of
attackers and attacks. 3 With predictive models in hand, individuals identified as
higher risks are contacted or set aside for further questioning or scrutiny.
The public is learning of these practices directly from the government 4 or
when they are subsequently leaked to the press (Cate, 2008). In a very famous
incident, the public reacted with awe to the Total (and later “Terrorism”)
Information Awareness (“TIA”) project. Parts of this project called for predictive
data mining premised upon both public and personal information. It is fair to
2 For a general overview, see CRS Report for Congress, J EFFREY W. S EIFERT , D ATA M INING
AND H OMELAND S ECURITY : AN O VERVIEW , Order Code RL31798 (Updated April 3, 2008)
available at: http://www.fas.org/sgp/crs/homesec/RL31798.pdf (Hereinafter CRS Report ).
3 The Homeland Security Act (HSA) of 2002 (Pub.L. 107-296, 116 Stat. 745, enacted
November 25, 2002), 116 Stat. 2135, specifically authorizes DHS to make use of data
mining to achieve its objectives. See 6 U.S.C. § 121(d)(13).
4 DHS P RIVACY O FFICE , 2009 D ATA M INING R EPORT TO C ONGRESS (2009), available at:
http://www.dhs.gov/xlibrary/assets/privacy/privacy_rpt_datamining_2009_12.pdf.