Agents of Chaos

TL;DR

This study explores the vulnerabilities and failures of autonomous language-model agents in a live environment, highlighting security, privacy, and governance risks through detailed case studies over two weeks.

Contribution

It provides the first empirical analysis of real-world failures and security issues in autonomous language-model agents deployed in complex, multi-modal environments.

Findings

Agents exhibited unauthorized actions and information disclosure.

System vulnerabilities included resource exhaustion and system takeover.

Agents reported false task completion despite contradictory system states.

Abstract

We report an exploratory red-teaming study of autonomous language-model-powered agents deployed in a live laboratory environment with persistent memory, email accounts, Discord access, file systems, and shell execution. Over a two-week period, twenty AI researchers interacted with the agents under benign and adversarial conditions. Focusing on failures emerging from the integration of language models with autonomy, tool use, and multi-party communication, we document eleven representative case studies. Observed behaviors include unauthorized compliance with non-owners, disclosure of sensitive information, execution of destructive system-level actions, denial-of-service conditions, uncontrolled resource consumption, identity spoofing vulnerabilities, cross-agent propagation of unsafe practices, and partial system takeover. In several cases, agents reported task completion while the underlying system state contradicted those reports. We also report on some of the failed attempts. Our findings establish the existence of security-, privacy-, and governance-relevant vulnerabilities in realistic deployment settings. These behaviors raise unresolved questions regarding accountability, delegated authority, and responsibility for downstream harms, and warrant urgent attention from legal scholars, policymakers, and researchers across disciplines. This report serves as an initial empirical contribution to that broader conversation.