Data Mining as Search: Theoretical Insights and Policy Responses - Discrimination and Privacy in the Information Society - page 317

various forms of ex ante judicial approval and supervision are required. If data
mining will be considered as a search, data mining analyses would be considered
an illegal search when carried out without sufficient judicial approval - approval
which is not currently sought.
The link between data mining practices and the concept of search can be made
on several levels - only one of which would be examined in this chapter. It could
be carried out on an intuitive level. It could also be carried out on a doctrinal level.
Finally, it could be carried out on a theoretical level. This chapter merely focuses
on the latter aspect. Yet before doing so, I hereby provide a few explanations
about the former two realms, and explain why I chose to set them aside for now.
On an intuitive level, data mining seems to invoke the notion of “searching”
and perhaps therefore, the legal implications of such terminology. The data mining
process calls for the substantial analyses of personal information pertaining to
specific individuals. In this process, computer programs work through a broad
array of datasets on their way to developing clusters, links, and other outputs.
Thereafter, the programs examine specific sets of personal data in real time in an
effort to establish whether they fit the predictive models previously constructed.
This is a process which will certainly be referred to as “searches” in laymen's
terms (Slobogin, 2007). Yet intuition is a fickle prospect. In many instances it
could be plainly wrong, as the public might be ill-informed regarding the true
meaning and implications of data mining - including its vast benefits. For that
reason, I set this discussion aside. Indeed, not all activities which are “searches” to
laymen are or should be considered as searches in the eyes of the law.
Linking data mining and searches will have real world implications and
therefore opens the door to an elaborate doctrinal analysis. When the law
recognizes searches as such, it moves to regulate them, limits their scope, and sets
systematic boundaries to assure the protection of rights. It is however unclear
whether under current case law and the existing concept of “search” as articulated
by the courts, data mining analyses constitute searches. In the US, for instance,
these steps are commonly discussed in the Fourth Amendment context, which
protects the people from unreasonable searches (Kerr, 2007). Whether current
Fourth Amendment doctrine will find data mining to be a “search” is a difficult
doctrinal question, which is beyond the scope of this chapter, but will probably be
answered negatively (Cate, 2008). Therefore, the starting point for this discussion
is that data mining analyses are not “searches.” The analysis set forth assumes that
data mining (or other forms of data analysis) is carried out while relying upon data
which was initially collected lawfully by either third parties and later passed on to
the government or directly (yet lawfully) by the government itself. With this
assumption in place, American law regarding searches generally assumes that
individuals have a very limited subsequent privacy interest (at least in terms of
“searching” and the Fourth Amendment) given the initial lawful collection of data
(Kerr, 2010). The point of data collection is where data subjects relinquish
control over the data and its future uses. To summarize, the governmental data
mining initiatives usually do not amount to breaches of constitutional rights; or, as
Daniel Solove succinctly states, “Data mining often falls between the crevices of