Saved in:
Bibliographic Details
Main Authors: Luo, Sha, Kim, Sang Jung, Duan, Zening, Chen, Kaiping
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2406.08222
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915577883262976
author Luo, Sha
Kim, Sang Jung
Duan, Zening
Chen, Kaiping
author_facet Luo, Sha
Kim, Sang Jung
Duan, Zening
Chen, Kaiping
contents Refusal behavior by Large Language Models is increasingly visible in content moderation, yet little is known about how refusals vary by the identity of the user making the request. This study investigates refusal as a sociotechnical outcome through a counterfactual persona design that varies gender identity--including male, female, non-binary, and transgender personas--while keeping the classification task and visual input constant. Focusing on a vision-language model (GPT-4V), we examine how identity-based language cues influence refusal in binary gender classification tasks. We find that transgender and non-binary personas experience significantly higher refusal rates, even in non-harmful contexts. Our findings also provide methodological implications for equity audits and content analysis using LLMs. Our findings underscore the importance of modeling identity-driven disparities and caution against uncritical use of AI systems for content coding. This study advances algorithmic fairness by reframing refusal as a communicative act that may unevenly regulate epistemic access and participation.
format Preprint
id arxiv_https___arxiv_org_abs_2406_08222
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Refusal as Silence: Gendered Disparities in Vision-Language Model Responses
Luo, Sha
Kim, Sang Jung
Duan, Zening
Chen, Kaiping
Computer Vision and Pattern Recognition
Artificial Intelligence
Computers and Society
Human-Computer Interaction
Refusal behavior by Large Language Models is increasingly visible in content moderation, yet little is known about how refusals vary by the identity of the user making the request. This study investigates refusal as a sociotechnical outcome through a counterfactual persona design that varies gender identity--including male, female, non-binary, and transgender personas--while keeping the classification task and visual input constant. Focusing on a vision-language model (GPT-4V), we examine how identity-based language cues influence refusal in binary gender classification tasks. We find that transgender and non-binary personas experience significantly higher refusal rates, even in non-harmful contexts. Our findings also provide methodological implications for equity audits and content analysis using LLMs. Our findings underscore the importance of modeling identity-driven disparities and caution against uncritical use of AI systems for content coding. This study advances algorithmic fairness by reframing refusal as a communicative act that may unevenly regulate epistemic access and participation.
title Refusal as Silence: Gendered Disparities in Vision-Language Model Responses
topic Computer Vision and Pattern Recognition
Artificial Intelligence
Computers and Society
Human-Computer Interaction
url https://arxiv.org/abs/2406.08222