Currently counter-surveillance focuses mainly upon either subverting the process of collection (e.g. wearing a rubber mask), or subverting the action (e.g. changing one's name to suggest a different ethnicity).
However, both of these approaches ignore a considerable part of the surveillance process, the analysis of the surveillance data. It is the analysis of surveillance data that allows for the building of the models that are used to sort people and objects. This work consists of two major components, the first of which is the delineation of the stages of surveillance, of which we argue that there are three: the collection, the analysis, and the action. In the next phase of this research we consider exactly what counter-surveillance means at the analysis stage. To do this, we explore how data may be manipulated to subvert analysis. We analyse three data-mining techniques, classification using both decision trees and support vector machines, and the development of association rules using the a priori algorithm. Each of these commonly used algorithms have unsuspected and significant vulnerabilities that may be easily exploited.
Everyone welcome!
The Surveillance Studies Centre
c/o David Murakami Wood
Department of Criminology
University of Ottawa
Canada