偏见、准确性与信任：基于性别多样性视角的大语言模型研究
摘要
大语言模型（LLM）在我们日常生活中越来越普遍，但关于LLM偏见的诸多担忧依然存在。本研究 examines how gender-diverse populations perceive bias, accuracy, and trustworthiness in LLMs, specifically ChatGPT. 通过对25名非二元/跨性别、男性和女性参与者进行深入访谈，我们探讨了性别化和中性提示如何影响模型响应，以及用户如何评估这些响应。我们的发现表明，性别化提示会引发更具身份特异性的响应，其中非二元参与者尤其容易受到轻蔑和刻板印象描绘的影响。 perceived accuracy across gender groups was consistent, with errors most noted in technical topics and creative tasks. Trustworthiness varied by gender, with men showing higher trust, especially in performance, and non-binary participants demonstrating higher performance-based trust. Additionally, participants suggested improving the LLMs by diversifying training data, ensuring equal depth in gendered responses, and incorporating clarifying questions. This research contributes to the CSCW/HCI field by highlighting the need for gender-diverse perspectives in LLM development in particular and AI in general, to foster more inclusive and trustworthy systems.
引用
@article{arxiv.2506.21898,
title = {Bias, Accuracy, and Trust: Gender-Diverse Perspectives on Large Language Models},
author = {Aimen Gaba and Emily Wall and Tejas Ramkumar Babu and Yuriy Brun and Kyle Hall and Cindy Xiong Bearfield},
journal= {arXiv preprint arXiv:2506.21898},
year = {2025}
}