@inproceedings{seo-etal-2025-taxonomy,
title = "Taxonomy of Comprehensive Safety for Clinical Agents",
author = "Seo, Jean and
Lee, Hyunkyung and
Kim, Gibaeg and
Han, Wooseok and
Yoo, Jaehyo and
Lim, Seungseop and
Shin, Kihun and
Yang, Eunho",
editor = "Potdar, Saloni and
Rojas-Barahona, Lina and
Montella, Sebastien",
booktitle = "Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing: Industry Track",
month = nov,
year = "2025",
address = "Suzhou (China)",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.emnlp-industry.124/",
pages = "1762--1779",
ISBN = "979-8-89176-333-3",
abstract = "Safety is a paramount concern in clinical chatbot applications, where inaccurate or harmful responses can lead to serious consequences. Existing methods{---}such as guardrails and tool-calling{---}often fall short in addressing the nuanced demands of the clinical domain. In this paper, we introduce TACOS(Taxonomy of Comprehensive Safety for Clinical Agents), a fine-grained, 21-class taxonomy that integrates safety filtering and tool selection into a single user intent classification step. TACOS covers a wide spectrum of clinical and non-clinical queries, explicitly modeling varying safety thresholds and external tool dependencies. To validate our taxonomy, we curate a TACOS-annotated dataset and perform extensive experiments. Our results demonstrate the value of a new taxonomy specialized for clinical agent settings, and reveal valuable insights about train data distribution and pretrained knowledge of base models."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="seo-etal-2025-taxonomy">
<titleInfo>
<title>Taxonomy of Comprehensive Safety for Clinical Agents</title>
</titleInfo>
<name type="personal">
<namePart type="given">Jean</namePart>
<namePart type="family">Seo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hyunkyung</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Gibaeg</namePart>
<namePart type="family">Kim</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wooseok</namePart>
<namePart type="family">Han</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jaehyo</namePart>
<namePart type="family">Yoo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Seungseop</namePart>
<namePart type="family">Lim</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kihun</namePart>
<namePart type="family">Shin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Eunho</namePart>
<namePart type="family">Yang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing: Industry Track</title>
</titleInfo>
<name type="personal">
<namePart type="given">Saloni</namePart>
<namePart type="family">Potdar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lina</namePart>
<namePart type="family">Rojas-Barahona</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sebastien</namePart>
<namePart type="family">Montella</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Suzhou (China)</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-333-3</identifier>
</relatedItem>
<abstract>Safety is a paramount concern in clinical chatbot applications, where inaccurate or harmful responses can lead to serious consequences. Existing methods—such as guardrails and tool-calling—often fall short in addressing the nuanced demands of the clinical domain. In this paper, we introduce TACOS(Taxonomy of Comprehensive Safety for Clinical Agents), a fine-grained, 21-class taxonomy that integrates safety filtering and tool selection into a single user intent classification step. TACOS covers a wide spectrum of clinical and non-clinical queries, explicitly modeling varying safety thresholds and external tool dependencies. To validate our taxonomy, we curate a TACOS-annotated dataset and perform extensive experiments. Our results demonstrate the value of a new taxonomy specialized for clinical agent settings, and reveal valuable insights about train data distribution and pretrained knowledge of base models.</abstract>
<identifier type="citekey">seo-etal-2025-taxonomy</identifier>
<location>
<url>https://aclanthology.org/2025.emnlp-industry.124/</url>
</location>
<part>
<date>2025-11</date>
<extent unit="page">
<start>1762</start>
<end>1779</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Taxonomy of Comprehensive Safety for Clinical Agents
%A Seo, Jean
%A Lee, Hyunkyung
%A Kim, Gibaeg
%A Han, Wooseok
%A Yoo, Jaehyo
%A Lim, Seungseop
%A Shin, Kihun
%A Yang, Eunho
%Y Potdar, Saloni
%Y Rojas-Barahona, Lina
%Y Montella, Sebastien
%S Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing: Industry Track
%D 2025
%8 November
%I Association for Computational Linguistics
%C Suzhou (China)
%@ 979-8-89176-333-3
%F seo-etal-2025-taxonomy
%X Safety is a paramount concern in clinical chatbot applications, where inaccurate or harmful responses can lead to serious consequences. Existing methods—such as guardrails and tool-calling—often fall short in addressing the nuanced demands of the clinical domain. In this paper, we introduce TACOS(Taxonomy of Comprehensive Safety for Clinical Agents), a fine-grained, 21-class taxonomy that integrates safety filtering and tool selection into a single user intent classification step. TACOS covers a wide spectrum of clinical and non-clinical queries, explicitly modeling varying safety thresholds and external tool dependencies. To validate our taxonomy, we curate a TACOS-annotated dataset and perform extensive experiments. Our results demonstrate the value of a new taxonomy specialized for clinical agent settings, and reveal valuable insights about train data distribution and pretrained knowledge of base models.
%U https://aclanthology.org/2025.emnlp-industry.124/
%P 1762-1779
Markdown (Informal)
[Taxonomy of Comprehensive Safety for Clinical Agents](https://aclanthology.org/2025.emnlp-industry.124/) (Seo et al., EMNLP 2025)
ACL
- Jean Seo, Hyunkyung Lee, Gibaeg Kim, Wooseok Han, Jaehyo Yoo, Seungseop Lim, Kihun Shin, and Eunho Yang. 2025. Taxonomy of Comprehensive Safety for Clinical Agents. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing: Industry Track, pages 1762–1779, Suzhou (China). Association for Computational Linguistics.