Who Gets the Mic? Investigating Gender Bias in the Speaker Assignment of a Speech-LLM

Dariia Puhach, Amir H. Payberah, Éva Székely
8/19/2025
cs.CLcs.AI

No official code found. Best implementation: daschablume/speech-gender-bias (0 stars, Apache-2.0, last commit 2025-08-21).

Abstract

Similar to text-based Large Language Models (LLMs), Speech-LLMs exhibit emergent abilities and context awareness. However, whether these similarities extend to gender bias remains an open question. This study proposes a methodology leveraging speaker assignment as an analytic tool for bias investigation. Unlike text-based models, which encode gendered associations implicitly, Speech-LLMs must produce a gendered voice, making speaker selection an explicit bias cue. We evaluate Bark, a Text-to-Speech (TTS) model, analyzing its default speaker assignments for textual prompts. If Bark's speaker selection systematically aligns with gendered associations, it may reveal patterns in its training data or model design. To test this, we construct two datasets: (i) Professions, containing gender-stereotyped occupations, and (ii) Gender-Colored Words, featuring gendered connotations. While Bark does not exhibit systematic bias, it demonstrates gender awareness and has some gender inclinations.

View on arXivView PDF

Code Implementations(1)

daschablume/speech-gender-biasCommunity68%
Apache-2.0

Repo for the paper "Who gets the mic? Investigating Gender Bias in the Speaker Assignment of a Speech-LLM"

00May 26, 20251 years agoApache-2.0

Similar Papers

Evaluating Gender Bias in Large Language Models

89% match1 repo2024

Acoustic-based Gender Differentiation in Speech-aware Language Models

88% match2025

Voice, Bias, and Coreference: An Interpretability Study of Gender in Speech Translation

88% match2025

Hire Me or Not? Examining Language Model's Behavior with Occupation Attributes

88% match1 repo2024

Addressing speaker gender bias in large scale speech translation systems

88% match2025

Assessing Gender Bias in LLMs: Comparing LLM Outputs with Human Perceptions and Official Statistics

87% match1 repo12024

Cite this paper

@article{puhach2025gets,
  title  = {Who Gets the Mic? Investigating Gender Bias in the Speaker Assignment of a Speech-LLM},
  author = {Dariia Puhach and Amir H. Payberah and Éva Székely},
  year   = {2025},
  eprint = {2508.13603},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2508.13603v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2508.13603.svg)](https://codeofpaper.com/paper/2508.13603)

Discussion