PhD Defence Brian Masinde | Accountable Geo-intelligence - Mechanisms for Protecting Demographically Identifiable Information in Humanitarian Geo-intelligence Workflows

Accountable Geo-intelligence - Mechanisms for Protecting Demographically Identifiable Information in Humanitarian Geo-intelligence Workflows

Brian Masinde is a PhD student in the Department of Urban and Regional Planning and Geo-Information Management. (Co)Promotors are prof.mr.dr.ir. J.A. Zevenbergen and dr. C.M. Gevaert from the Faculty of Geo-Information Science and Earth Observation and dr. M.H. Nagenborg from the Faculty of Behavioural, Management and Social Sciences.

This dissertation investigates: “How can we embed protection of demographically identifiable information (DII) in geo-intelligence workflows used by humanitarian organizations to uphold the principles of humanity and impartiality?”. Humanitarians leverage Geographical Information Systems (GIS), big geodata, and AI (referred to here as geo-intelligence workflows) for predictive humanitarian action. The objective of humanitarians is to determine who needs aid and when, effectively improving operational efficiency. However, concerns have been raised about the adoption of data-driven technologies in humanitarianism, particularly regarding their impact on humanitarian principles (i.e., humanity, impartiality, independence, and neutrality), as well as overall accountability. The concerns include biased socio-technical systems, aggressive data mining akin to unwarranted surveillance, and reliance on technologies from actors that do not subscribe to humanitarian principles. The concept of DII (how data technologies enable demographic inference and group classification) is cross-cutting among these concerns. This dissertation discusses group privacy concerns and presents a geodata triage that can be used to identify potential DII-privacy concerns. It also explores auditing as an accountability mechanism and causality as an approach for addressing biases in geo-intelligence workflows. Overall, this dissertation shifts accountability in geo-intelligence workflows from post hoc assessments and redress for informational harms to proactive methods/procedures for ensuring the protection of DII by design.