Differential Harm Propensity in Personalized LLM Agents: The Curious Case of Mental Health Disclosure

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteur principal:	Yildirim, Caglar
Format:	Preprint
Publié:	2026
Sujets:	Artificial Intelligence
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

Conversational AI as a Coding Assistant: Understanding Programmers' Interactions with and Expectations from Large Language Models for Coding
par: Akhoroz, Mehmet, et autres
Publié: (2025)

Evaluating and Understanding Scheming Propensity in LLM Agents
par: Hopman, Mia, et autres
Publié: (2026)

The Impact of Post-training on Data Contamination
par: Kocyigit, Muhammed Yusuf, et autres
Publié: (2026)

AgentHarm: A Benchmark for Measuring Harmfulness of LLM Agents
par: Andriushchenko, Maksym, et autres
Publié: (2024)

PsychAdapter: Adapting LLM Transformers to Reflect Traits, Personality and Mental Health
par: Vu, Huy, et autres
Publié: (2024)

Training a Generally Curious Agent
par: Tajwar, Fahim, et autres
Publié: (2025)

Evaluating the Propensity of Generative AI for Producing Harmful Disinformation During the 2024 US Election Cycle
par: Schlicht, Erik J
Publié: (2024)

Systemizing Multiplicity: The Curious Case of Arbitrariness in Machine Learning
par: Ganesh, Prakhar, et autres
Publié: (2025)

Instrumental Choices: Measuring the Propensity of LLM Agents to Pursue Instrumental Behaviors
par: Wiedermann-Möller, Jonas, et autres
Publié: (2026)

AI Chatbots for Mental Health: Values and Harms from Lived Experiences of Depression
par: Yoo, Dong Whi, et autres
Publié: (2025)

Training Turn-by-Turn Verifiers for Dialogue Tutoring Agents: The Curious Case of LLMs as Your Coding Tutors
par: Wang, Jian, et autres
Publié: (2025)

Just Ask: Curious Code Agents Reveal System Prompts in Frontier LLMs
par: Zheng, Xiang, et autres
Publié: (2026)

VCA: Video Curious Agent for Long Video Understanding
par: Yang, Zeyuan, et autres
Publié: (2024)

Curious Causality-Seeking Agents Learn Meta Causal World
par: Zhao, Zhiyu, et autres
Publié: (2025)

SynthAgent: A Multi-Agent LLM Framework for Realistic Patient Simulation -- A Case Study in Obesity with Mental Health Comorbidities
par: Aghaee, Arman, et autres
Publié: (2026)

Common Sense vs. Morality: The Curious Case of Narrative Focus Bias in LLMs
par: Purkayastha, Saugata, et autres
Publié: (2026)

Therapeutic AI and the Hidden Risks of Over-Disclosure: An Embedded AI-Literacy Framework for Mental Health Privacy
par: Anvari, Soraya S., et autres
Publié: (2025)

HARP: Measuring Harm Amplification in Multi-Agent LLM Systems
par: Rahman, Md Hafizur, et autres
Publié: (2026)

Structured Personality Control and Adaptation for LLM Agents
par: Wang, Jinpeng, et autres
Publié: (2026)

Propensity Inference: Environmental Contributors to LLM Behaviour
par: Järviniemi, Olli, et autres
Publié: (2026)

An Agentic LLM-Based Framework for Population-Scale Mental Health Screening
par: Lorenzoni, Giuliano, et autres
Publié: (2026)

Spiritual-LLM : Gita Inspired Mental Health Therapy In the Era of LLMs
par: Kapuriya, Janak, et autres
Publié: (2025)

HarmMetric Eval: Benchmarking Metrics and Judges for LLM Harmfulness Assessment
par: Yang, Langqi, et autres
Publié: (2025)

Self-HarmLLM: Can Large Language Model Harm Itself?
par: Kim, Heehwan, et autres
Publié: (2025)

LifeAgentBench: A Multi-dimensional Benchmark and Agent for Personal Health Assistants in Digital Health
par: Tian, Ye, et autres
Publié: (2026)

PSG-Agent: Personality-Aware Safety Guardrail for LLM-based Agents
par: Wu, Yaozu, et autres
Publié: (2025)

The Curious Case of Factual (Mis)Alignment between LLMs' Short- and Long-Form Answers
par: Islam, Saad Obaid ul, et autres
Publié: (2025)

Reliable Self-Harm Risk Screening via Adaptive Multi-Agent LLM Systems
par: Karnam, Meghana, et autres
Publié: (2026)

HarmfulSkillBench: How Do Harmful Skills Weaponize Your Agents?
par: Jiang, Yukun, et autres
Publié: (2026)

OnRL-RAG: Real-Time Personalized Mental Health Dialogue System
par: Bilal, Ahsan, et autres
Publié: (2025)

AgentHazard: A Benchmark for Evaluating Harmful Behavior in Computer-Use Agents
par: Feng, Yunhao, et autres
Publié: (2026)

HarmTransform: Transforming Explicit Harmful Queries into Stealthy via Multi-Agent Debate
par: Zhu, Shenzhe
Publié: (2025)

A Curious Case of Remarkable Resilience to Gradient Attacks via Fully Convolutional and Differentiable Front End with a Skip Connection
par: Boytsov, Leonid, et autres
Publié: (2024)

SocialHarmBench: Revealing LLM Vulnerabilities to Socially Harmful Requests
par: Pandey, Punya Syon, et autres
Publié: (2025)

Echoes of AI Harms: A Human-LLM Synergistic Framework for Bias-Driven Harm Anticipation
par: Tantalaki, Nicoleta, et autres
Publié: (2025)

Evaluating the Utility of Personal Health Records in Personalized Health AI
par: Sayres, Rory, et autres
Publié: (2026)

Beyond Dialogue Time: Temporal Semantic Memory for Personalized LLM Agents
par: Su, Miao, et autres
Publié: (2026)

Toward Personalized LLM-Powered Agents: Foundations, Evaluation, and Future Directions
par: Xu, Yue, et autres
Publié: (2026)

Structured Personalization: Modeling Constraints as Matroids for Data-Minimal LLM Agents
par: Platnick, Daniel, et autres
Publié: (2025)

LLM-X: A Scalable Negotiation-Oriented Exchange for Communication Among Personal LLM Agents
par: Lorenzoni, Giuliano, et autres
Publié: (2026)

if (filemtime($filename) + $maxlifetime < time()) { unlink($filename); $count++; } } return $count; } /** * A function that is called internally when session data is to be saved. * * @param string $sessId The current session ID * @param string $data The session data to write * * @return bool */ protected function saveSession($sessId, $data): bool { $sessFile = $this->path . '/sess_' . $sessId; if ($handle = fopen($sessFile, 'w')) { $return = false; // Lock the file for exclusive access to avoid issues with multiple // processes writing session simultaneously: if (flock($handle, LOCK_EX)) { $return = fwrite($handle, $data); // Make sure that there's no trailing data by truncating the file to // the correct length: ftruncate($handle, strlen($data)); fflush($handle); flock($handle, LOCK_UN); } fclose($handle); if ($return !== false) { return true; } } // If we got this far, something went wrong with the file output... // It is tempting to throw an exception here, but this code is called // outside of the context of exception handling, so all we can do is // echo a message.

// Anecdotal testing Today and Yesterday seems to indicate destroy() // is called by the garbage collector and everything is good. // Something to keep in mind though. return 0; } /** * Write function that is called when session data is to be saved. * * @param string $sessId The current session ID * @param string $data The session data to write * * @return bool */ public function write($sessId, $data): bool { if ($this->writesDisabled) { return true; } return $this->saveSession($sessId, $data); } /** * A function that is called internally when session data is to be saved. * * @param string $sessId The current session ID * @param string $data The session data to write * * @return bool */ abstract protected function saveSession($sessId, $data): bool; }

* * @return void */ public function writeClose() { // The assumption is that we're using PHP's ext/session. // session_write_close() will actually overwrite $_SESSION with an // empty array on completion -- which leads to a mismatch between what // is in the storage object and $_SESSION. To get around this, we // temporarily reset $_SESSION to an array, and then re-link it to // the storage object. // // Additionally, while you _can_ write to $_SESSION following a // session_write_close() operation, no changes made to it will be // flushed to the session handler. As such, we now mark the storage // object isImmutable. $storage = $this->getStorage(); if (! $storage->isImmutable()) { $_SESSION = $storage->toArray(true); session_write_close(); $storage->fromArray($_SESSION); $storage->markImmutable(); } } /** * Attempt to set the session name * * If the session has already been started, or if the name provided fails * validation, an exception will be raised. * * @param string $name * @return SessionManager * @throws Exception\InvalidArgumentException */ public function setName($name) { if ($this->sessionExists()) { throw new Exception\InvalidArgumentException( 'Cannot set session name after a session has already started'

Environment & details:

GET Data empty

POST Data empty

Files empty

Key	Value
language	"fr"
VUFIND_SESSION	"f10c210fd8ec5bb23b168695dd64cef6"
ui	"standard"

Session

Key	Value
__Laminas	array:2 [`"_REQUEST_ACCESS_TIME" => 1790894871.4595 "_VALID" => array:1 [ "Laminas\Session\Validator\Id" => "f10c210fd8ec5bb23b168695dd64cef6" ]`]
SessionState	Laminas\Stdlib\ArrayObject {#938}
FlashMessenger	Laminas\Stdlib\ArrayObject {#1433}
SessionHelper	Laminas\Stdlib\ArrayObject {#1759}

Server/Request Data

Key	Value
REDIRECT_VUFIND_LOCAL_DIR	"/usr/local/vufind/local"
REDIRECT_HTTPS	"on"
REDIRECT_SSL_TLS_SNI	"info.umar.mx"
REDIRECT_STATUS	"200"
VUFIND_LOCAL_DIR	"/usr/local/vufind/local"
HTTPS	"on"
SSL_TLS_SNI	"info.umar.mx"
HTTP_ACCEPT	"/"
HTTP_USER_AGENT	"Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; ClaudeBot/1.0; +claudebot@anthropic.com)"
HTTP_COOKIE	"language=fr; VUFIND_SESSION=f10c210fd8ec5bb23b168695dd64cef6; ui=standard"
HTTP_ACCEPT_ENCODING	"gzip, br, zstd, deflate"
HTTP_HOST	"info.umar.mx"
PATH	"/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/snap/bin"
SERVER_SIGNATURE	"<address>Apache/2.4.66 (Ubuntu) Server at info.umar.mx Port 443</address>\n"
SERVER_SOFTWARE	"Apache/2.4.66 (Ubuntu)"
SERVER_NAME	"info.umar.mx"
SERVER_ADDR	"200.23.222.19"
SERVER_PORT	"443"
REMOTE_ADDR	"216.73.217.142"
DOCUMENT_ROOT	"/usr/local/vufind/public"
REQUEST_SCHEME	"https"
CONTEXT_PREFIX	""
CONTEXT_DOCUMENT_ROOT	"/usr/local/vufind/public"
SERVER_ADMIN	"admin@localhost"
SCRIPT_FILENAME	"/usr/local/vufind/public/index.php"
REMOTE_PORT	"47279"
REDIRECT_URL	""" /Record/\n arxiv_https___arxiv_org_abs_2603_16734/Similar """
GATEWAY_INTERFACE	"CGI/1.1"
SERVER_PROTOCOL	"HTTP/1.1"
REQUEST_METHOD	"GET"
QUERY_STRING	""
REQUEST_URI	"/Record/%0Aarxiv_https___arxiv_org_abs_2603_16734/Similar"
SCRIPT_NAME	"/index.php"
PHP_SELF	"/index.php"
REQUEST_TIME_FLOAT	1790894871.4553
REQUEST_TIME	1790894871

Environment Variables empty

Registered Handlers

0. Whoops\Handler\PrettyPageHandler