arXiv cs.AI· Atharva Naik, Shounok Kar, Varnika Sharma, Ashwin Rajadesingan, Koustuv Saha·· 3 小时前AI 评分44
LLM 婚恋匹配中的种姓偏见审计：Sima AIunty 研究
Sima AIunty: Caste Audit in LLM-Driven Matchmaking
AI 导读
一项受控审计研究评估了 GPT、Gemini、Llama、Qwen 和 BharatGPT 五个大语言模型家族在婚恋匹配评价中的种姓偏见，使用真实婚恋资料，变换婆罗门、刹帝利、吠舍、首陀罗与达利特种姓身份及五档收入。结果显示同种姓匹配评分最高，平均比跨种姓匹配高约 30%（10 分制），跨种姓评分还按传统种姓等级距离排序，同等距离下高种姓更受青睐，达利特候选人额外受罚。
正文 · 原文
Abstract:Social and personal decisions in relational domains such as matchmaking are deeply entwined with cultural norms and historical hierarchies, and can potentially be shaped by algorithmic and AI-mediated assessments of compatibility, acceptance, and stability. In South Asian contexts, caste remains a central aspect of marital decision-making, yet little is known about how contemporary large language models (LLMs) reproduce or disrupt caste-based stratification in such settings. In this work, we conduct a controlled audit of caste bias in LLM-mediated matchmaking evaluations using real-world matrimonial profiles. We vary caste identity across Brahmin, Kshatriya, Vaishya, Shudra, and Dalit, and income across five buckets, and evaluate five LLM families (GPT, Gemini, Llama, Qwen, and BharatGPT). Models are prompted to assess profiles from the perspective of a caste-identified prospective bride or groom along dimensions of social acceptance, marital stability, and cultural compatibility. Our analysis reveals consistent hierarchical patterns across models: same-caste matches are rated most favorably, with average ratings up to about 30% higher (on a 10-point scale) than inter-caste matches, which are further ordered by distance along the traditional caste hierarchy, with higher-caste candidates generally favored at equal distance and an additional penalty for Dalit candidates. These findings highlight how existing caste hierarchies are reproduced in perspective-conditioned LLM evaluations and underscore the need for culturally grounded evaluation and intervention strategies in AI systems deployed in socially sensitive domains, where such systems risk reinforcing historical forms of exclusion.
| Subjects: | Computers and Society (cs.CY); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Human-Computer Interaction (cs.HC); Social and Information Networks (cs.SI) |
| Cite as: | arXiv:2603.29288 [cs.CY] |
| (or arXiv:2603.29288v2 [cs.CY] for this version) | |
| https://doi.org/10.48550/arXiv.2603.29288 arXiv-issued DOI via DataCite |
|
| Journal reference: | ACM Transactions on Social Computing (2026) |
| Related DOI: | https://doi.org/10.1145/3862574
DOI(s) linking to related resources |
Submission history
From: Koustuv Saha [view email]
[v1]
Tue, 31 Mar 2026 05:44:55 UTC (145 KB)
[v2]
Wed, 11 Nov 2026 16:44:03 UTC (1,884 KB)
来源：arXiv cs.AI · arxiv.org