An Exploration of Automatic Speech Recognition Within a Nordic Context.

Automatic speech recognition (ASR) has become a familiar input component for voice user faces, such as texting, preparing structured documents, searching, and voice commands. Inclusive ASR is a premise for a sustainable HCI that supports all types of languages and users. ASR is quite robust for majority languages but still needs to be adequate for smaller languages. Moreover, specific user groups cannot use ASR due to, e.g., speech impediments or accents. This paper discusses the possibilities and limitations of ASR in its current form within a Nordic context raising the following questions: i) What is the state of ASR in terms of usability for smaller and minority languages, and ii) How is ASR discussed regarding future development and inclusion? The current research discusses such topics within the lens of ability-based participatory design established on exploratory data collection. Among other findings, the paper stresses that small and minority languages are neglected in ASR development from an international perspective. Consequently, these languages risk exclusion from the digital development of language technology.