OS-Harm: A Benchmark for Measuring Safety of Computer Use Agents

2025
arxiv.org

Thomas Kuntz, Agatha Duzan, Hao Zhao, Francesco Croce, Zico Kolter, Nicolas Flammarion, Maksym Andriushchenko

ELLIS Authors

Francesco Croce
Member
ELLIS Institute Finland
Aalto University
Nicolas Flammarion
Member
Swiss Federal Technology Institute of Lausanne (EPFL)
Maksym Andriushchenko
Member
ELLIS Institute Tübingen
Max Planck Institute for Intelligent Systems (MPI-IS)
Tübingen AI Center

Author Locations

No location data available for the ELLIS authors of this paper.

Home
Research
PhD & Postdoc Program Sites Research Programs Jobs ELLIS PhD Award Projects Building on ELLIS Conference Contributions
Members
Become a Member Members List Become a Fellow Fellows List
About
Board Organisation ELLIS Open Letter ELLIS Position Paper Contact For Media ELLIS FAQ Partnerships
News
Events
Jobs Become a Member For Media Contact Intranet
X LinkedIn Bluesky Facebook
European Laboratory for Learning and Intelligent Systems
ELLIS Edge Newsletter
Join the 6,000+ people who get the monthly newsletter filled with the latest news, jobs, events and insights from the ELLIS Network.
X LinkedIn Bluesky Facebook