Technical and Professional Communicators as Advocates of Linguistic Justice in the Design of Speech Technologies

Abstract

Despite claims of being a revolutionary technology, research demonstrates that speech technologies, broadly, and automatic speech recognition systems, specifically continue to demonstrate listening (recognition) bias against languages and dialects spoken by people of color, foreign speakers of English, and marginalized communities. Yet despite evidence of bias, listening devices are increasingly being used in US schools, prisons, courts of law, and workplaces such as call centers in India and the Philippines: spaces all disproportionately represented by people of color and foreign English speakers. The paper reframes the on-going conversation around linguistic representation in speech technologies as an urgent linguistic justice issue by highlighting the sociopolitical contexts in which these devices are used. I argue that technical communicators—both researchers and practitioners—are in an ideal position to advocate for a more socially just design of speech devices and to assess and mitigate potential harms to marginalized communities with varying language backgrounds.