Taking Ethics, Fairness, and Bias Seriously in Machine Learning for Disaster Risk Management

Abstract: This paper highlights an important, if under-examined, set of questions about the deployment of machine learning technologies in the field of disaster risk management (DRM). While emerging tools show promising capacity to support scientific efforts to better understand and mitigate the threats posed by disasters and climate change, our field must undertake a much more careful assessment of the potential negative impacts that machine learning technologies may create. We also argue that attention to these issues in the context of machine learning affords the opportunity to have discussions about potential ethics, bias, and fairness concerns within disaster data more broadly. In what follows, we first describe some of the uses and potential benefits of machine-learning technology in disaster risk management. We then draw on research from other fields to speculate about potential negative impacts. Finally, we outline a research agenda for how our disaster risk management can begin to take these issues seriously and ensure that deployments of machine-learning tools are conducted in a responsible and beneficial manner.