Bias in the Ear of the Listener: Assessing Sensitivity in Audio Language Models Across Linguistic, Demographic, and Positional Variations

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Wei, Sheng-Lun, Liao, Yu-Ling, Chang, Yen-Hua, Huang, Hen-Hsen, Chen, Hsin-Hsi
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866908802359492608
author Wei, Sheng-Lun
Liao, Yu-Ling
Chang, Yen-Hua
Huang, Hen-Hsen
Chen, Hsin-Hsi
author_facet Wei, Sheng-Lun
Liao, Yu-Ling
Chang, Yen-Hua
Huang, Hen-Hsen
Chen, Hsin-Hsi
contents This work presents the first systematic investigation of speech bias in multilingual MLLMs. We construct and release the BiasInEar dataset, a speech-augmented benchmark based on Global MMLU Lite, spanning English, Chinese, and Korean, balanced by gender and accent, and totaling 70.8 hours ($\approx$4,249 minutes) of speech with 11,200 questions. Using four complementary metrics (accuracy, entropy, APES, and Fleiss' $κ$), we evaluate nine representative models under linguistic (language and accent), demographic (gender), and structural (option order) perturbations. Our findings reveal that MLLMs are relatively robust to demographic factors but highly sensitive to language and option order, suggesting that speech can amplify existing structural biases. Moreover, architectural design and reasoning strategy substantially affect robustness across languages. Overall, this study establishes a unified framework for assessing fairness and robustness in speech-integrated LLMs, bridging the gap between text- and speech-based evaluation. The resources can be found at https://github.com/ntunlplab/BiasInEar.
format Preprint
id arxiv_https___arxiv_org_abs_2602_01030
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Bias in the Ear of the Listener: Assessing Sensitivity in Audio Language Models Across Linguistic, Demographic, and Positional Variations
Wei, Sheng-Lun
Liao, Yu-Ling
Chang, Yen-Hua
Huang, Hen-Hsen
Chen, Hsin-Hsi
Computation and Language
Sound
Audio and Speech Processing
This work presents the first systematic investigation of speech bias in multilingual MLLMs. We construct and release the BiasInEar dataset, a speech-augmented benchmark based on Global MMLU Lite, spanning English, Chinese, and Korean, balanced by gender and accent, and totaling 70.8 hours ($\approx$4,249 minutes) of speech with 11,200 questions. Using four complementary metrics (accuracy, entropy, APES, and Fleiss' $κ$), we evaluate nine representative models under linguistic (language and accent), demographic (gender), and structural (option order) perturbations. Our findings reveal that MLLMs are relatively robust to demographic factors but highly sensitive to language and option order, suggesting that speech can amplify existing structural biases. Moreover, architectural design and reasoning strategy substantially affect robustness across languages. Overall, this study establishes a unified framework for assessing fairness and robustness in speech-integrated LLMs, bridging the gap between text- and speech-based evaluation. The resources can be found at https://github.com/ntunlplab/BiasInEar.
title Bias in the Ear of the Listener: Assessing Sensitivity in Audio Language Models Across Linguistic, Demographic, and Positional Variations
topic Computation and Language
Sound
Audio and Speech Processing
url https://arxiv.org/abs/2602.01030