On Preserving Privacy Whilst Integrating Data in Connected Information Systems

摘要

Currently information systems collect, process and disseminate a huge amount of data that usually contains privacy sensitive information. With the advent of cloud computing and big data paradigms we witness that data often propagates from its origin (i.e., data subjects) and goes through a number of data processing units. Along its path the data is quite likely processed and integrated with other information about the subject in every data processing unit. One can foresee that at a point in the chain of data processors a data processor may infer more privacy sensitive information about the data subject than the data subject agrees with. This can lead to privacy breaches if inadequate privacy preserving measures are in place. In this contribution we firstly elaborate on the necessity of a so-called forward policy propagation mechanism in a chain of data processors. This mechanism allows propagating the privacy preferences of individuals along with their data items. These data items, if appropriate measures are not taken, can potentially be merged with other information about the subject in a way that privacy sensitiveinformation about the user is derived. Knowing the privacy preferences, a data processing unit can ensure the enforcement of user privacy preferences on the processed data by, for example, re-anonymising it according to the privacy preferences. We further argue that, on ground of on-going initiatives and trends, one also needs to inform the upstream data processors whose information items are used to infer more privacy sensitive information about the user than it is allowed. This will enable the upstream data processors to revisit their data sharing policies in light of the occurred privacy breaches. Therefore we further propose a so-called backward event propagation mechanism to report data breach events towards upstream data processors. We also briefly touch upon the legal implications and legislative trends/gaps in this regard and the impacts of such a feedback mechanism on Open Data initiatives.