@inproceedings{nghiem-etal-2026-bias,
title = "Bias in the Tails: How Name-conditioned Evaluative Framing in Resume Summaries Destabilizes {LLM}-based Hiring",
author = "Nghiem, Huy and
Nguyen-Le, Phuong-Anh and
Ho, Sy-Tuyen and
Daum{\'e} III, Hal",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-main.1623/",
pages = "36889--36933",
ISBN = "979-8-89176-459-0",
abstract = "Research has documented LLMs' name-based bias in hiring and salary recommendations. In this paper, we instead consider a setting where LLMs generate candidate summaries for downstream assessment. In a large-scale controlled study, we analyze nearly one million resume summaries produced by 4 models under systematic race{--}gender name perturbations, using synthetic O*NET-grounded resumes and real-world job postings. By decomposing each summary into resume-grounded factual content and evaluative framing, we find that factual content remains largely stable, while evaluative language exhibits subtle name-conditioned variation concentrated in the extremes of the distribution, especially in open-source models. Our LLM-judged hiring simulation demonstrates how evaluative summary transforms directional harm into symmetric instability that might evade conventional fairness audits in LLM-mediated pipelines, highlighting a potential pathway for LLM-to-LLM automation bias."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="nghiem-etal-2026-bias">
<titleInfo>
<title>Bias in the Tails: How Name-conditioned Evaluative Framing in Resume Summaries Destabilizes LLM-based Hiring</title>
</titleInfo>
<name type="personal">
<namePart type="given">Huy</namePart>
<namePart type="family">Nghiem</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Phuong-Anh</namePart>
<namePart type="family">Nguyen-Le</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sy-Tuyen</namePart>
<namePart type="family">Ho</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hal</namePart>
<namePart type="family">Daumé III</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-459-0</identifier>
</relatedItem>
<abstract>Research has documented LLMs’ name-based bias in hiring and salary recommendations. In this paper, we instead consider a setting where LLMs generate candidate summaries for downstream assessment. In a large-scale controlled study, we analyze nearly one million resume summaries produced by 4 models under systematic race–gender name perturbations, using synthetic O*NET-grounded resumes and real-world job postings. By decomposing each summary into resume-grounded factual content and evaluative framing, we find that factual content remains largely stable, while evaluative language exhibits subtle name-conditioned variation concentrated in the extremes of the distribution, especially in open-source models. Our LLM-judged hiring simulation demonstrates how evaluative summary transforms directional harm into symmetric instability that might evade conventional fairness audits in LLM-mediated pipelines, highlighting a potential pathway for LLM-to-LLM automation bias.</abstract>
<identifier type="citekey">nghiem-etal-2026-bias</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.1623/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>36889</start>
<end>36933</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Bias in the Tails: How Name-conditioned Evaluative Framing in Resume Summaries Destabilizes LLM-based Hiring
%A Nghiem, Huy
%A Nguyen-Le, Phuong-Anh
%A Ho, Sy-Tuyen
%A Daumé III, Hal
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F nghiem-etal-2026-bias
%X Research has documented LLMs’ name-based bias in hiring and salary recommendations. In this paper, we instead consider a setting where LLMs generate candidate summaries for downstream assessment. In a large-scale controlled study, we analyze nearly one million resume summaries produced by 4 models under systematic race–gender name perturbations, using synthetic O*NET-grounded resumes and real-world job postings. By decomposing each summary into resume-grounded factual content and evaluative framing, we find that factual content remains largely stable, while evaluative language exhibits subtle name-conditioned variation concentrated in the extremes of the distribution, especially in open-source models. Our LLM-judged hiring simulation demonstrates how evaluative summary transforms directional harm into symmetric instability that might evade conventional fairness audits in LLM-mediated pipelines, highlighting a potential pathway for LLM-to-LLM automation bias.
%U https://aclanthology.org/2026.emnlp-main.1623/
%P 36889-36933
Markdown (Informal)
[Bias in the Tails: How Name-conditioned Evaluative Framing in Resume Summaries Destabilizes LLM-based Hiring](https://aclanthology.org/2026.emnlp-main.1623/) (Nghiem et al., EMNLP 2026)
ACL