EAGER: Collaborative: A Research Agenda to Explore Privacy in Small Data Applications

点击翻译按钮获取中文摘要

One of the crucial ideas behind Privacy by Design (PbD) is that privacy should be taken into consideration in the process of design, not merely after-the-fact, as so often happens. Yet, this ideal has failed to gain widespread practical traction, challenged, in part, by the lack of developed methodologies and also because of privacy's conceptual complexity, which hampers its operationalization. This project addresses both challenges simultaneously, seeking (i) to demonstrate how a robust operationalization of privacy can lead to meaningful PbD and (ii) to contribute methodological insight by engaging with ongoing research and development in the area of small data applications, namely, systems that advance wellness and personal productivity by utilizing digital traces from individuals' day-to-day activities, such as e-mail, grocery shopping, TV watching, transportation, mobile devices, and so forth. Adopting the definition of privacy as contextual integrity, the project will focus on selected small-data applications currently "on the drawing board" in PI Deborah Estrin's Small Data lab. With these design cases, the project rises to one of the PbD challenges, namely consideration of privacy early on in development and, as a research enterprise, its primary aim is to uncover more and less productive methodological approaches for doing so, resulting in system characteristics well correlated with privacy requirements. At the same time, the project will provide invaluable insight into how to operationalize contextual integrity, which conceives of privacy as appropriate flow of personal information, modeling appropriate flow as conformance with context-specific informational norms, which, prescribe (and prohibit) information flows according to three parameters: actors (subject, sender, recipient), information types, and transmission principles (functional constraints on flow). Adopting contextual integrity as an operational definition means that researchers will assess privacy properties by carefully mapping data flows, and evaluating these flows in terms of the context of application and use. The project also extends past work on formal representations of informational norms by demonstrating how they may be integrated into design practices. In addition to its substantive contributions this project embodies an innovative collaborative model -- a novel pairing of a computer scientist, Deborah Estrin (Cornell), with a philosopher, Helen Nissenbaum (NYU), in an equal partnership to forge technologies that embody meaningful privacy.