AI & ML
Colosseum: Auditing Collusion in Cooperative Multi-Agent Systems
arXiv cs.MA0 views
arXiv:2602.15198v3 Announce Type: replace Abstract: Multi-agent systems, where LLM agents communicate through free-form language, enable sophisticated coordination for solving complex cooperative tasks. This surfaces a unique safety problem when a group of agents forms a coalition and colludes to pursue secondary goals and degrade the joint objective. In this paper, we present Colosseum, a framework for auditing LLM agents' collusive behavior in multi-agent settings. We ground how agents cooperate through a formal multi-agent decision-making framework and measure action-based collusive behavior in actions via regret relative to the cooperative optimum and compare it with communication-based collusive behavior. Colosseum enables audits of LLM agents for collusion under benign settings, different coalition objectives, persuasion tactics, and network topologies. We then introduce a new behavioral probe by creating secret communication channels between agents, showing that most out-of-the-box models exhibit a propensity to collude under this probe, which we term emergent collusion. Furthermore, we discover ``collusion on paper'' when agents plan to collude in text but often pick non-collusive actions. Colosseum provides a new way to audit collusion in cooperative multi-agent systems while presenting observations about how collusion emerges, what affects collusion efficacy, and which strategies may mitigate it.
Source:https://arxiv.org/abs/2602.15198PreviousUncertainty, bias and the institution bootstrapping problemNextDisagreement-Guided Agentic Data Curation for Multimodal Safety
Related posts

Reducing Rollout Bottlenecks in Reinforcement Learning with NVIDIA Dynamo
0NVIDIA Developer Blog
Open-Source Roundup: Video Reverse Engineering, Isometric Illustrations & Agent Automation
0dev.to AI

I Replaced My 5-Year-Old Self-Built PC — and Ran a 125B Local LLM on an RTX 5070 Ti
0dev.to AI
Yahoo Scout Signals Why Search Visibility Audits Must Go Beyond Google
0dev.to AI
Google brings AI conversations directly to the main search page
0dev.to AI
Catching Silent AI Agent Failures in Production: Metrics, Drift and Hallucinated Actions
0dev.to Security