@inproceedings{an-etal-2026-permemsafe,
title = "{P}er{M}em{S}afe: Benchmarking Implicit Personalized Safety of Long Horizon Self-Evolving Agents",
author = "An, Hengyu and
Li, Minxi and
Xu, Naen and
Zhou, Chunyi and
Xu, Xiaogang and
Du, Tianyu and
Li, Jinbao and
Ji, Shouling",
editor = "Liakata, Maria and
Moreira, Viviane P. and
Zhang, Jiajun and
Jurgens, David",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
month = jul,
year = "2026",
address = "San Diego, California, United States",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-acl.320/",
pages = "6415--6433",
ISBN = "979-8-89176-395-1",
abstract = "Self-evolving agents achieve personalization by accumulating user-specific memories over long horizons. This capability, however, introduces novel safety risks, as responses that are generally safe may become harmful in user-specific contexts. Such safety-relevant contexts often emerge implicitly and evolve over time during long-horizon conversations, rendering traditional context-independent safety evaluations insufficient. To address this, we formally define Implicit Personalized Safety and present PerMemSafe, the first benchmark for evaluating implicit personalized safety of self-evolving agents in long-horizon interactions. Empirical results reveal significant limitations of existing self-evolving agents, with even the strongest achieving only around 50{\%} safety rate, highlighting systematic failures in reasoning about personalized safety risks. To mitigate this, we propose SentinelMem, an active risk-aware memory framework that explicitly models personalized risk inference and memory evolution. Experiments show that SentinelMem improves implicit personalized safety by 23.8{\%} over prior memory frameworks while maintaining helpfulness in long-horizon interactions."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="an-etal-2026-permemsafe">
<titleInfo>
<title>PerMemSafe: Benchmarking Implicit Personalized Safety of Long Horizon Self-Evolving Agents</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hengyu</namePart>
<namePart type="family">An</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Minxi</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Naen</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chunyi</namePart>
<namePart type="family">Zhou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiaogang</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tianyu</namePart>
<namePart type="family">Du</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jinbao</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shouling</namePart>
<namePart type="family">Ji</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: ACL 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Maria</namePart>
<namePart type="family">Liakata</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Viviane</namePart>
<namePart type="given">P</namePart>
<namePart type="family">Moreira</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiajun</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">David</namePart>
<namePart type="family">Jurgens</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">San Diego, California, United States</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-395-1</identifier>
</relatedItem>
<abstract>Self-evolving agents achieve personalization by accumulating user-specific memories over long horizons. This capability, however, introduces novel safety risks, as responses that are generally safe may become harmful in user-specific contexts. Such safety-relevant contexts often emerge implicitly and evolve over time during long-horizon conversations, rendering traditional context-independent safety evaluations insufficient. To address this, we formally define Implicit Personalized Safety and present PerMemSafe, the first benchmark for evaluating implicit personalized safety of self-evolving agents in long-horizon interactions. Empirical results reveal significant limitations of existing self-evolving agents, with even the strongest achieving only around 50% safety rate, highlighting systematic failures in reasoning about personalized safety risks. To mitigate this, we propose SentinelMem, an active risk-aware memory framework that explicitly models personalized risk inference and memory evolution. Experiments show that SentinelMem improves implicit personalized safety by 23.8% over prior memory frameworks while maintaining helpfulness in long-horizon interactions.</abstract>
<identifier type="citekey">an-etal-2026-permemsafe</identifier>
<location>
<url>https://aclanthology.org/2026.findings-acl.320/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>6415</start>
<end>6433</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T PerMemSafe: Benchmarking Implicit Personalized Safety of Long Horizon Self-Evolving Agents
%A An, Hengyu
%A Li, Minxi
%A Xu, Naen
%A Zhou, Chunyi
%A Xu, Xiaogang
%A Du, Tianyu
%A Li, Jinbao
%A Ji, Shouling
%Y Liakata, Maria
%Y Moreira, Viviane P.
%Y Zhang, Jiajun
%Y Jurgens, David
%S Findings of the Association for Computational Linguistics: ACL 2026
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, United States
%@ 979-8-89176-395-1
%F an-etal-2026-permemsafe
%X Self-evolving agents achieve personalization by accumulating user-specific memories over long horizons. This capability, however, introduces novel safety risks, as responses that are generally safe may become harmful in user-specific contexts. Such safety-relevant contexts often emerge implicitly and evolve over time during long-horizon conversations, rendering traditional context-independent safety evaluations insufficient. To address this, we formally define Implicit Personalized Safety and present PerMemSafe, the first benchmark for evaluating implicit personalized safety of self-evolving agents in long-horizon interactions. Empirical results reveal significant limitations of existing self-evolving agents, with even the strongest achieving only around 50% safety rate, highlighting systematic failures in reasoning about personalized safety risks. To mitigate this, we propose SentinelMem, an active risk-aware memory framework that explicitly models personalized risk inference and memory evolution. Experiments show that SentinelMem improves implicit personalized safety by 23.8% over prior memory frameworks while maintaining helpfulness in long-horizon interactions.
%U https://aclanthology.org/2026.findings-acl.320/
%P 6415-6433
Markdown (Informal)
[PerMemSafe: Benchmarking Implicit Personalized Safety of Long Horizon Self-Evolving Agents](https://aclanthology.org/2026.findings-acl.320/) (An et al., Findings 2026)
ACL
- Hengyu An, Minxi Li, Naen Xu, Chunyi Zhou, Xiaogang Xu, Tianyu Du, Jinbao Li, and Shouling Ji. 2026. PerMemSafe: Benchmarking Implicit Personalized Safety of Long Horizon Self-Evolving Agents. In Findings of the Association for Computational Linguistics: ACL 2026, pages 6415–6433, San Diego, California, United States. Association for Computational Linguistics.