Ethical considerations on the use of machine translation and crowdsourcing in cascading crises

Abstract

When a sudden-onset emergency occurs, the language needs of those affected and those involved in the relief efforts cannot be foreseen. Provided that access to online communication is still available, it is not unlikely that many involved in the crisis will resort to language technologies such as machine translation and initiatives such as crowdsourcing to assist in the urgent need for multilingual communication. This may be done in an attempt to understand the key messages from official bodies, or relief organisations, when there is a lack of professional translators to assist in the multilingual communication process. This approach machine translation and crowdsourcing was successfully used in a previous crisis, i.e. the 2010 earthquake in Haiti. However, the use of technologies cannot be taken for granted. Even if they are supposedly used for good, a number of ethical issues should be given consideration before using these technologies, when using them, and in the aftermath of a crisis. In this chapter, we describe those issues by having a closer look at potential crisis translation workflows which rely on machine translation and crowdsourcing.