Voice AI and Hermeneutical Injustice at the Border

Abstract

I argue that Voice AI systems perpetuates hermeneutical injustice by preventing marginalised individuals from equitably participating in defining the social roles that they need to make sense of their own experiences, or render themselves intelligible to others. I draw upon voice biometrics at the German border used to assess asylum seeker applications, and accent alteration for call centre workers to illustrate how the relevant concepts of the legitimate asylum seeker or competent call centre worker are unduly influenced by the respective Voice AI. Ultimately, the hermeneutical injustices reflect the epistemic hegemony of AI steeped in dominant Western knowledge structures that reinforce post-colonial dynamics of identity erasure and disruption.