Perng and Buscher Uncertainty and Transparency Short Paper – Ethical, Legal and Social Issues Proceedings of the ISCRAM 2015 Conference - Kristiansand, May 24-27 Palen, Büscher, Comes & Hughes, eds. Uncertainty and Transparency: Augmenting Modelling and Prediction for Crisis Response Sung-Yueh Perng Maynooth University, Ireland sung-yueh.perng@nuim.ie Monika Buscher Lancaster University m.buscher@lancaster.ac.uk ABSTRACT Emergencies are characterised by uncertainty. This motivates the design of information systems that model and predict complex natural, material or human processes to support understanding and reduce uncertainty through prediction. The correspondence between system models and reality, however, is also governed by uncertainties, and designers have developed methods to render ‘the world’ transparent in ways that can inform, fine-tune and validate models. Additionally, people experience uncertainties in their use of simulation and prediction systems. This is a major obstacle to effective utilisation. We discuss ethically and socially motivated demands for transparency. Keywords Uncertainty, transparency, reasoning, modelling prediction, collaboration INTRODUCTION When emergencies occur, they are characterised by their ‘un-ness’ – they are unexpected, unprecedented and unplanned for in their specific unfolding (Crichton 2003, cited in McMaster and Baber, 2008, p. 6). As these authors identify, the ‘un-ness’ can result in a number of uncertainties particularly when multiple agencies are involved. These include uncertainties about the nature of the crisis (fire or explosion?), details (location? access?), its cause (gas leak or terrorist attack?), available quality of information (source, relevance,accuracy?) and response goals and strategies. Modelling techniques have been developed to address these issues (Turoff, Hiltz, Bañuls, and Van Den Eede, 2013; Haynes, Jermusyk, and Ritter, 2014). Increasingly, the capability of generating and processing diverse, continuous and even live data feeds can augment situation awareness and decision-making. A recently developed BRIDGE project concept, for example, incorporates sensor data captured by drones, allowing dynamic analysis of environmental and meteorological data in an emergency caused by fire, which can be further interpreted by 3D models to estimate the threat to buildings, the immediate environment, victims and first responders and to recommend response strategies and necessary equipment (Steinhäusler, 2015). Predictive analytics can further augment crisis detection and response by detecting abnormal patterns with trained algorithms. While it is true that new capability for collecting a wider range of real-time data and feeding analyses to responders can support more detailed and constantly updated situational awareness as well as improved sense- and decision-making, these potentials generate new uncertainties, as well as ethical and legal issues. Whether data, analysis and prediction can be effectively utilised, depends critically on human and social factors. To chart ELSI in relation to uncertainty, simulation and prediction, this paper