Cover Image

When Benign Inputs Lead to Severe Harms: Eliciting Unsafe Unintended Behaviors of Computer-Use Agents

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Jones, Jaylen, Zhang, Zhehao, Ning, Yuting, Fosler-Lussier, Eric, St-Charles, Pierre-Luc, Bengio, Yoshua, Song, Dawn, Su, Yu, Sun, Huan
Format:	Preprint
Published:	2026
Subjects:	Computation and Language Artificial Intelligence Cryptography and Security
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

A Multi-Aspect Framework for Counter Narrative Evaluation using Large Language Models
by: Jones, Jaylen, et al.
Published: (2024)

RedTeamCUA: Realistic Adversarial Testing of Computer-Use Agents in Hybrid Web-OS Environments
by: Liao, Zeyi, et al.
Published: (2025)

When Actions Go Off-Task: Detecting and Correcting Misaligned Actions in Computer-Use Agents
by: Ning, Yuting, et al.
Published: (2026)

Music on the Move
by: Fosler-Lussier, Danielle
Published: (2020)

Active Attacks: Red-teaming LLMs via Adaptive Environments
by: Yun, Taeyoung, et al.
Published: (2025)

Improving Transducer-Based Spoken Language Understanding with Self-Conditioned CTC and Knowledge Transfer
by: Sunder, Vishal, et al.
Published: (2025)

Latent Fusion Jailbreak: Blending Harmful and Harmless Representations to Elicit Unsafe LLM Outputs
by: Xing, Wenpeng, et al.
Published: (2025)

Improving Speech Recognition Error Prediction for Modern and Off-the-shelf Speech Recognizers
by: Serai, Prashant, et al.
Published: (2024)

End-to-End Diarization utilizing Attractor Deep Clustering
by: Palzer, David, et al.
Published: (2025)

Improving Neural Diarization through Speaker Attribute Attractors and Local Dependency Modeling
by: Palzer, David, et al.
Published: (2025)

Machine learning and information theory concepts towards an AI Mathematician
by: Bengio, Yoshua, et al.
Published: (2024)

AmpleGCG-Plus: A Strong Generative Model of Adversarial Suffixes to Jailbreak LLMs with Higher Success Rates in Fewer Attempts
by: Kumar, Vishal, et al.
Published: (2024)

On the Proactive Generation of Unsafe Images From Text-To-Image Models Using Benign Prompts
by: Wu, Yixin, et al.
Published: (2023)

Baking Symmetry into GFlowNets
by: Ma, George, et al.
Published: (2024)

VISTA: Verification In Sequential Turn-based Assessment
by: Lewis, Ashley, et al.
Published: (2025)

Beyond Length: Context-Aware Expansion and Independence as Developmentally Sensitive Evaluation in Child Utterances
by: Chun, Jiyun, et al.
Published: (2026)

Can a Bayesian Oracle Prevent Harm from an Agent?
by: Bengio, Yoshua, et al.
Published: (2024)

Do LLM Decoders Listen Fairly? Benchmarking How Language Model Priors Shape Bias in Speech Recognition
by: Ginjala, Srishti, et al.
Published: (2026)

Noticing the Watcher: LLM Agents Can Infer CoT Monitoring from Blocking Feedback
by: Jiralerspong, Thomas, et al.
Published: (2026)

Hijack-GAN: Unintended-Use of Pretrained, Black-Box GANs
by: Wang, Hui-Po, et al.
Published: (2020)

Unintended Harms of Value-Aligned LLMs: Psychological and Empirical Insights
by: Choi, Sooyung, et al.
Published: (2025)

Eliciting Harmful Capabilities by Fine-Tuning On Safeguarded Outputs
by: Kaunismaa, Jackson, et al.
Published: (2026)

Beyond Visual Safety: Jailbreaking Multimodal Large Language Models for Harmful Image Generation via Semantic-Agnostic Inputs
by: Yu, Mingyu, et al.
Published: (2026)

The Surprising Harmfulness of Benign Overfitting for Adversarial Robustness
by: Hao, Yifan, et al.
Published: (2024)

When Good Sounds Go Adversarial: Jailbreaking Audio-Language Models with Benign Inputs
by: Dingeto, Hiskias, et al.
Published: (2025)

HarmAug: Effective Data Augmentation for Knowledge Distillation of Safety Guard Models
by: Lee, Seanie, et al.
Published: (2024)

On Generalization for Generative Flow Networks
by: Krichel, Anas, et al.
Published: (2024)

Interventional Causal Representation Learning
by: Ahuja, Kartik, et al.
Published: (2022)

A Complexity-Based Theory of Compositionality
by: Elmoznino, Eric, et al.
Published: (2024)

Visual symbolic mechanisms: Emergent symbol processing in vision language models
by: Assouel, Rim, et al.
Published: (2025)

Relative Trajectory Balance is equivalent to Trust-PCL
by: Deleu, Tristan, et al.
Published: (2025)

Fast Monte Carlo Tree Diffusion: 100x Speedup via Parallel Sparse Planning
by: Yoon, Jaesik, et al.
Published: (2025)

In-Context Parametric Inference: Point or Distribution Estimators?
by: Mittal, Sarthak, et al.
Published: (2025)

Eliciting Uncertainty in Chain-of-Thought to Mitigate Bias against Forecasting Harmful User Behaviors
by: Sicilia, Anthony, et al.
Published: (2024)

Balancing Efficiency and Safety: How and When Algorithmic Management Induces Gig Workers' Unsafe Behavior
by: Yanghao Zhu, et al.
Published: (2026)

GFlowNet Foundations
by: Bengio, Yoshua, et al.
Published: (2021)

When 2D Tasks Meet 1D Serialization: On Serialization Friction in Structured Tasks
by: Lo, Chung-Hsiang, et al.
Published: (2026)

Can Safety Fine-Tuning Be More Principled? Lessons Learned from Cybersecurity
by: Williams-King, David, et al.
Published: (2025)

RL, but don't do anything I wouldn't do
by: Cohen, Michael K., et al.
Published: (2024)

LLMGuard: Guarding Against Unsafe LLM Behavior
by: Goyal, Shubh, et al.
Published: (2024)