Beyond the binary: Limitations and possibilities of gender-related speech technology research

TL;DR

The paper addresses the problem that gender-related speech research at Interspeech has been scarce and often relies on underspecified, binary definitions of gender. It conducts a systematic review of 107 Interspeech papers from 2013–2023 to assess how gender, sex, and perceived gender are defined and used, revealing pervasive lack of definition and binary framing. The authors highlight the risks of excluding marginalised groups and discuss inclusive dataset practices, terminology clarity, and research motivations, offering concrete guidelines for more responsible work. The study underscores the need for explicit terminology, transparent data practices, and ethical consideration to avoid reinforcing biases and to expand representation in speech technology.

Abstract

This paper presents a review of 107 research papers relating to speech and sex or gender in ISCA Interspeech publications between 2013 and 2023. We note the scarcity of work on this topic and find that terminology, particularly the word gender, is used in ways that are underspecified and often out of step with the prevailing view in social sciences that gender is socially constructed and is a spectrum as opposed to a binary category. We draw attention to the potential problems that this can cause for already marginalised groups, and suggest some questions for researchers to ask themselves when undertaking work on speech and gender.

Figures (2)

• Figure 1: Breakdown of the question Is a definition for gender/sex provided?, per year.
• Figure 2: Breakdown of the question Does the definition of gender/sex go beyond the binary?, per year.