Transparency in Data Mining: From Theory to Practice - Discrimination and Privacy in the Information Society - page 302

relevant. Individuals will contribute to this project altruistically. Others will do so
as a hobby or as a part of their academic research. Yet others might do so as
means to contribute to a community which might be emerging (Citron, 2008).
Thus, it is fair to assume that a sufficient number of individuals will strive to
review and contribute to these policies and governmental initiatives.
Current transparency regulation does not reflect any aspects of this theory.
Generally, government does not enable any meaningful feedback of the prediction
process. The most practical segment for implementing such policy is where its
absence is most noticeable - with regard to the computer code charged with
running the analysis. However, rather than allowing experts to review and comment
on it, the government provides almost no insights as to the codes inner workings.
While applying this rationale into policy is important, it could be substituted at
times by providing information to a selected group of experts These experts will
assist the government with feedback on predictive modeling projects without
disclosing the information further. Shifting to this limited form of disclosure might
be called for given the strong arguments against full disclosure (such as, that
disclosing source code will compromise trade secrets and the overall success of
the prediction scheme).
Autonomy as Control over Personal Data
In the process of predictive modeling, a requirement for transparency flows
directly from the rights of those individuals whose personal information was used
throughout the process - the “data subjects.”
The basic premise leading to this aspect of transparency is the notion of control
individuals have over their personal information (Westin, 1967; Lessig, 1999).
This theoretical notion has been broadly accepted in the EU, 18 while only partially
recognized in the US. This concept could be understood as an extension of the
individual's autonomy. It was translated into several concrete principles that after
several transitions formulated the “Fair Information Practices,” or FIPs. 19
“Openness” or “Transparency” was central to FIPs from their earliest stage
(Reidenberg, 1995). In FIP's current version, this notion is encapsulated in the
principle of “Notice.” “Notice” commonly refers to informing individuals that
personal information about them is being collected, and its subsequent uses. The
analysis of personal information is an essential part of the predictive processes
here discussed. Thus, recognizing the principle of “Notice” should lead
governments to shed additional light on the data mining processes as far as they
pertain to personal information.
18 Directive 95/46/EC on the protection of individuals with regard to the processing of
personal data and on the free movement of such data, Official Journal of the European
Communities of 23 November 1995 No L. 281 p. 31 (Hereinafter EU Data Protection
Directive ); D ANIEL J. S OLOVE & P AUL M. S CHWARTZ , I NFORMATION P RIVACY L AW
(Aspen Publishers, 2006), 35-8.
19 O RGANIZATION FOR E CONOMIC C O - OPERATION AND D EVELOPMENT (OECD), G UIDELINES
ON THE P ROTECTION OF P RIVACY AND T RANSBORDER F LOWS OF P ERSONAL D ATA (1980).