Despite the fact that variation is a fundamental characteristic of natural language, automatic speech recognition systems perform systematically worse on non-standardised and marginalised language varieties. In this paper we use the lens of language policy to analyse how current practices in training and testing ASR systems in industry lead to the data bias giving rise to these systematic error differences. We believe that this is a useful perspective for speech and language technology practitioners to understand the origins and harms of algorithmic bias, and how they can mitigate it. We also propose a re-framing of language resources as (public) infrastructure which should not solely be designed for markets, but for, and with meaningful cooperation of, speech communities.

Algorithmic oppression is an urgent and persistent problem in speech and language technologies. Considering power relations embedded in datasets before compiling or using them to train or test speech and language technologies is essential to designing less harmful, more just technologies. This paper presents a reflective exercise to recognise and challenge gaps and the power relations they reveal in speech and language datasets by applying principles of Data Feminism and Design Justice, and building on work on dataset documentation and sociolinguistics.

Commercial Automatic Speech Recognition (ASR) systems tend to show systemic predictive bias for marginalised speaker/user groups. We highlight the need for an interdisciplinary and context-sensitive approach to documenting this bias incorporating perspectives and methods from sociolinguistics, speech & language technology and human-computer interaction in the context of a case study. We argue evaluation of ASR systems should be disaggregated by speaker group, include qualitative error analysis, and consider user experience in a broader sociolinguistic and social context.

Multi-sentence questions (MSQs) are sequences of questions connected by relations which, unlike sequences of standalone questions, need to be answered as a unit. Following Rhetorical Structure Theory (RST), we recognise that different “question discourse relations” between the subparts of MSQs reflect different speaker intents, and consequently elicit different answering strategies. Correctly identifying these relations is therefore a crucial step in automatically answering MSQs. We identify five different types of MSQs in English, and define five novel relations to describe them. We extract over 162,000 MSQs from Stack Exchange to enable future research. Finally, we implement a high-precision baseline classifier based on surface features.