Ethics, Automated Processes, Machine Translation, and Crises

Abstract

AbstractDeploying technologies in support of translation/interpreting during crises in multilingual settings poses serious deontological and ethical challenges. Most arise from ethical concerns around the adoption of technologies that can be only partially controlled. We discuss automated translation processes in relation to four dimensions of preparedness, crowdsourcing and data mining, local vs global crises, and multimodal demands of communication. We start by considering automation processes in which MT is embedded as a tool to support crisis communication, we consider ethical risks pertaining reliance on and understanding of MT potential. We then focus on the ethical complexity of multimodal processes of communication that hinge on crowdsourcing practices, that collate users’ data, and that complement other automation processes. We move on to correlate these explicit ethical dimensions, with successful applications of MT engines to respond to local and global crises, and reflect on the ethical need to enshrine these in other practices that make translation into a risk reduction tool. We eventually zoom in on areas that are yet to exploit fully current automation processes, yet have already encountered ethical dilemmas when delivering information in multimodal format. We look at ways in which current automation processes can be successfully exploited, while we also warn to revise practices in which too much is expected by MT and automation processes thus heightening rather than reducing risks when communicating in multilingual crises. We conclude the chapter by connecting our ethical considerations on the role of MT and automation to debates around linguistic equality and social justice.KeywordsMultimodal crisis communicationHuman-machine integrationTranslation as risk reductionCrisis translation