arXiv cs.AI· Nicol\`o Pagan, Michael Andreas Ebling, Stefania Ionescu, Anik\'o Hann\'ak·· 4 小时前AI 评分54
arXiv 论文发现 AI 评估模型偏好特定 AI 撰写的求职信而与候选人质量无关
AI evaluators favor job applications written by certain AI models, independent of candidate quality
AI 导读
一项 arXiv 研究构建多模型审计框架，用 11 个 LLM 写作 agent 配 9 个评估 agent、10 份职位描述和每职位 50 名候选人，检验模型选择是否独立于资质影响招聘结果。研究发现评估模型存在全局一致的偏好层级，本地部署模型系统性吃亏，且有自我偏好；最不利配对下 25 名合格候选人中约 1.8 人纯因模型错配被筛掉，偏好源于写作风格线索与评估模型对这些线索的一致奖惩。
正文 · 原文
Abstract:Job applicants increasingly use large language models (LLMs) to write cover letters, while employers deploy LLMs to screen them. In this all-AI setting, we ask whether the \emph{choice of model} affects hiring outcomes independently of candidate qualifications. We introduce a multi-model auditing framework pairing eleven LLM writer agents with nine API-hosted evaluator agents across ten job descriptions and fifty candidates per job. Four findings emerge. First, we identify a generative preference hierarchy: a consistent global ranking of writer models by evaluator acceptance, in which locally hosted models are systematically disadvantaged regardless of their license, compounded by self-preference bias where evaluators favour their own outputs. Second, evaluator models differ substantially in how much bias they introduce based on applicants' tool choice, offering actionable guidance for fairer screening design. Third, this hierarchy has measurable competitive consequences: in the most adverse pairing, approximately two (1.8) out of twenty-five qualified candidates are filtered out purely due to model mismatch. Fourth, a common mechanism explains the patterns above: writer preference and self-preference trace to a combination of how much each writer leans on specific stylistic cues and how strongly evaluators, often unanimously, reward or penalise those same cues. A further check finds little evidence that an unmeasured "writer fingerprint" remains once these same features are accounted for. We discuss implications for algorithmic fairness in employment, regulation of AI-assisted screening, and the design of auditing frameworks that can be reapplied as the model landscape evolves.
| Subjects: | Computers and Society (cs.CY); Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA) |
| Cite as: | arXiv:2610.41866 [cs.CY] |
| (or arXiv:2610.41866v1 [cs.CY] for this version) | |
| https://doi.org/10.48550/arXiv.2610.41866 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Nicolò Pagan Mr. [view email]
[v1]
Fri, 30 Oct 2026 15:18:29 UTC (10,985 KB)
来源：arXiv cs.AI · arxiv.org