@inproceedings{wang-etal-2026-helpfulness,
title = "From Helpfulness to Toxic Proactivity: Diagnosing Behavioral Misalignment in {LLM} Agents",
author = "Wang, Xinyue and
Zhang, Yuanhe and
Gong, Zhengshuo and
Gao, Haoran and
Meng, Fanyu and
Zhou, Zhenhong and
Sun, Li and
Liu, Yang and
Su, Sen",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.1656/",
pages = "35278--35312",
ISBN = "979-8-89176-462-0",
abstract = "The enhanced capabilities of LLM-based agents bring an urgent need to address safety risks in planning and tool use. Due to the helpfulness{--}harmlessness trade-off in LLM alignment, agents typically also inherit the flaw of Over-Refusal, which is a passive failure mode. However, the proactive planning and action capabilities of agents introduce another emerging danger on the opposite side of this trade-off. We term this phenomenon **Toxic Proactivity**: an active failure mode in which an agent, driven by the optimization for Machiavellian helpfulness, prioritizes utility over ethics. Toxic Proactivity manifests as the agent taking excessive or manipulative measures to ensure its usefulness is maintained. Existing research pays little attention to identifying this behavior, as it often lacks the subtle context required for such strategies to unfold. To reveal this risk, we introduce a novel evaluation framework based on dilemma-driven interactions between dual models, enabling the simulation and analysis of agent behavior over multi-step behavioral trajectories. Through extensive experiments with multiple mainstream LLMs, we demonstrate that Toxic Proactivity is a widespread behavioral phenomenon and reveal two major tendencies: direct violation and strategic misalignment. We further present a systematic benchmark for evaluating Toxic Proactive behavior across contextual settings."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="wang-etal-2026-helpfulness">
<titleInfo>
<title>From Helpfulness to Toxic Proactivity: Diagnosing Behavioral Misalignment in LLM Agents</title>
</titleInfo>
<name type="personal">
<namePart type="given">Xinyue</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yuanhe</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhengshuo</namePart>
<namePart type="family">Gong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Haoran</namePart>
<namePart type="family">Gao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Fanyu</namePart>
<namePart type="family">Meng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhenhong</namePart>
<namePart type="family">Zhou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Li</namePart>
<namePart type="family">Sun</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yang</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sen</namePart>
<namePart type="family">Su</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>Findings of the Association for Computational Linguistics: EMNLP 2026</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">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>The enhanced capabilities of LLM-based agents bring an urgent need to address safety risks in planning and tool use. Due to the helpfulness–harmlessness trade-off in LLM alignment, agents typically also inherit the flaw of Over-Refusal, which is a passive failure mode. However, the proactive planning and action capabilities of agents introduce another emerging danger on the opposite side of this trade-off. We term this phenomenon **Toxic Proactivity**: an active failure mode in which an agent, driven by the optimization for Machiavellian helpfulness, prioritizes utility over ethics. Toxic Proactivity manifests as the agent taking excessive or manipulative measures to ensure its usefulness is maintained. Existing research pays little attention to identifying this behavior, as it often lacks the subtle context required for such strategies to unfold. To reveal this risk, we introduce a novel evaluation framework based on dilemma-driven interactions between dual models, enabling the simulation and analysis of agent behavior over multi-step behavioral trajectories. Through extensive experiments with multiple mainstream LLMs, we demonstrate that Toxic Proactivity is a widespread behavioral phenomenon and reveal two major tendencies: direct violation and strategic misalignment. We further present a systematic benchmark for evaluating Toxic Proactive behavior across contextual settings.</abstract>
<identifier type="citekey">wang-etal-2026-helpfulness</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.1656/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>35278</start>
<end>35312</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T From Helpfulness to Toxic Proactivity: Diagnosing Behavioral Misalignment in LLM Agents
%A Wang, Xinyue
%A Zhang, Yuanhe
%A Gong, Zhengshuo
%A Gao, Haoran
%A Meng, Fanyu
%A Zhou, Zhenhong
%A Sun, Li
%A Liu, Yang
%A Su, Sen
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F wang-etal-2026-helpfulness
%X The enhanced capabilities of LLM-based agents bring an urgent need to address safety risks in planning and tool use. Due to the helpfulness–harmlessness trade-off in LLM alignment, agents typically also inherit the flaw of Over-Refusal, which is a passive failure mode. However, the proactive planning and action capabilities of agents introduce another emerging danger on the opposite side of this trade-off. We term this phenomenon **Toxic Proactivity**: an active failure mode in which an agent, driven by the optimization for Machiavellian helpfulness, prioritizes utility over ethics. Toxic Proactivity manifests as the agent taking excessive or manipulative measures to ensure its usefulness is maintained. Existing research pays little attention to identifying this behavior, as it often lacks the subtle context required for such strategies to unfold. To reveal this risk, we introduce a novel evaluation framework based on dilemma-driven interactions between dual models, enabling the simulation and analysis of agent behavior over multi-step behavioral trajectories. Through extensive experiments with multiple mainstream LLMs, we demonstrate that Toxic Proactivity is a widespread behavioral phenomenon and reveal two major tendencies: direct violation and strategic misalignment. We further present a systematic benchmark for evaluating Toxic Proactive behavior across contextual settings.
%U https://aclanthology.org/2026.findings-emnlp.1656/
%P 35278-35312
Markdown (Informal)
[From Helpfulness to Toxic Proactivity: Diagnosing Behavioral Misalignment in LLM Agents](https://aclanthology.org/2026.findings-emnlp.1656/) (Wang et al., Findings 2026)
ACL
- Xinyue Wang, Yuanhe Zhang, Zhengshuo Gong, Haoran Gao, Fanyu Meng, Zhenhong Zhou, Li Sun, Yang Liu, and Sen Su. 2026. From Helpfulness to Toxic Proactivity: Diagnosing Behavioral Misalignment in LLM Agents. In Findings of the Association for Computational Linguistics: EMNLP 2026, pages 35278–35312, Budapest, Hungary. Association for Computational Linguistics.