<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="chang-etal-2026-beyond-semantic">
    <titleInfo>
        <title>Beyond Semantic Accuracy: Consequence-Aware Evaluation for Safety-Critical Language Understanding</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Yujing</namePart>
        <namePart type="family">Chang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Duc-thinh</namePart>
        <namePart type="family">Pham</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Van-Phat</namePart>
        <namePart type="family">Thai</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Chunyao</namePart>
        <namePart type="family">Ma</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Yash</namePart>
        <namePart type="family">Guleria</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Pham</namePart>
        <namePart type="given">Nhut</namePart>
        <namePart type="family">Huy</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Sameer</namePart>
        <namePart type="family">Alam</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>Can language models be trusted in safety-critical operations? In such settings, strong performance on semantic metrics does not guarantee operational reliability: a misread altitude, a dropped execution condition, or a confused callsign may score well under standard F1 yet carry sharply asymmetric operational consequences. We study this problem in air traffic control (ATC), where controller-pilot communication demands near-zero error tolerance, and use consequence-aware evaluation to test whether semantic scores misstate operational reliability. The framework is instantiated in a controlled diagnostic ATC benchmark grounded in aviation standards and feedback from 40 air traffic controllers across three countries. Evaluating 8 models, we uncover a systematic semantic-safety gap: conventional scores give substantially higher performance estimates than consequence-aware evaluation, even for models that appear reliable under standard metrics. Risk-aware fine-tuning narrows but does not close this gap, showing that consequence-aware evaluation is a necessary complement to standard NLP metrics before any real safety-critical deployment claim.</abstract>
    <identifier type="citekey">chang-etal-2026-beyond-semantic</identifier>
    <location>
        <url>https://aclanthology.org/2026.emnlp-main.2141/</url>
    </location>
    <part>
        <date>2026-10</date>
        <extent unit="page">
            <start>48673</start>
            <end>48708</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T Beyond Semantic Accuracy: Consequence-Aware Evaluation for Safety-Critical Language Understanding
%A Chang, Yujing
%A Pham, Duc-thinh
%A Thai, Van-Phat
%A Ma, Chunyao
%A Guleria, Yash
%A Huy, Pham Nhut
%A Alam, Sameer
%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 chang-etal-2026-beyond-semantic
%X Can language models be trusted in safety-critical operations? In such settings, strong performance on semantic metrics does not guarantee operational reliability: a misread altitude, a dropped execution condition, or a confused callsign may score well under standard F1 yet carry sharply asymmetric operational consequences. We study this problem in air traffic control (ATC), where controller-pilot communication demands near-zero error tolerance, and use consequence-aware evaluation to test whether semantic scores misstate operational reliability. The framework is instantiated in a controlled diagnostic ATC benchmark grounded in aviation standards and feedback from 40 air traffic controllers across three countries. Evaluating 8 models, we uncover a systematic semantic-safety gap: conventional scores give substantially higher performance estimates than consequence-aware evaluation, even for models that appear reliable under standard metrics. Risk-aware fine-tuning narrows but does not close this gap, showing that consequence-aware evaluation is a necessary complement to standard NLP metrics before any real safety-critical deployment claim.
%U https://aclanthology.org/2026.emnlp-main.2141/
%P 48673-48708

Markdown (Informal)

[Beyond Semantic Accuracy: Consequence-Aware Evaluation for Safety-Critical Language Understanding](https://aclanthology.org/2026.emnlp-main.2141/) (Chang et al., EMNLP 2026)

• Beyond Semantic Accuracy: Consequence-Aware Evaluation for Safety-Critical Language Understanding (Chang et al., EMNLP 2026)

ACL

• Yujing Chang, Duc-thinh Pham, Van-Phat Thai, Chunyao Ma, Yash Guleria, Pham Nhut Huy, and Sameer Alam. 2026. Beyond Semantic Accuracy: Consequence-Aware Evaluation for Safety-Critical Language Understanding. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 48673–48708, Budapest, Hungary. Association for Computational Linguistics.