Database Reference
In-Depth Information
correlation the affected individual falls within. When provided with such
information, the individual could rest assured there was no identification error.
She might further understand that the process was not random. Yet without
understanding the inner workings leading to this outcome, these results might still
appear arbitrary. Therefore, providing limited insights into the governmental
actions at stage (c) might be insufficient for restoring autonomy and dignity.
In view of the above and to further empower relevant individuals and provide
meaningful feedback, this theory calls for expanding transparency beyond stage
(c) of the data mining process. It probably requires that data mining would be an
interpretable process. It might even call for assuring a causation theory was found
to explain all actions taken. With such additional disclosure, individuals can
obtain sufficient insight to the process and how it relates to their lives. This theory
could also be understood to call for transparency in other stages of the prediction
process. It could call for the measures described above as part of stage (d). In
other words, to assure dignity and promote autonomy, the individual should
receive assurances as to the precision, effectiveness and lack of discrimination in
the process. The information provided through the feedback loop can promote
these objectives.
Furthermore, this theory could also justify transparency in stage (b).
Information regarding the use of data mining algorithms is essential to allow the
affected individual to retain autonomy and dignity. With information regarding
these important steps of the process, the mere correlations used (in which an
individual was implicated) can be understood as part of a broader picture. This
will prove helpful in understanding that targeting was not arbitrary, and perhaps
even in challenging its findings. It should be noted, however, that this segment of
the argument is relatively weak.
Thus far, this section of the analysis has shown that this autonomy-based
rationale provides a powerful argument for transparency in a broad variety of
segments along the data mining process. However, this theory also includes a
central flaw - it provides a transparency justification for merely a small segment
of the population - those adversely impacted by the relevant predictive practices.
Only such individuals face the potential of autonomy-based harms and are thus
entitled to autonomy-based remedies.
Yet one can argue convincingly that if the government must disclose such
information to a limited population segment, it should already provide it to the
entire public. This argument flows from acknowledging that the information
regarding the data mining practices vested with the few will make its way to the
entire population anyway. In today's information age, it is quite common that
disclosure to a limited group of disgruntled individuals quickly leads to spreading
such knowledge to the entire public. Those adversely affected will provide their
information online (and if stigma may attach, will do so anonymously). With time,
the pieces of the puzzle will come together and a full picture would emerge in the
public realm. For that reason, government should initially go ahead and provide
such information to all.
At this point, some might argue that disclosing these governmental practices to
the affected few will not lead to a broad understanding of what the government is
Search WWH ::




Custom Search