▶ Watch videoMatthew Lee
Collusion in multi-agent systems
Defining unwanted coordination, understanding risks to oversight, and studying covert communication between AI agents.
New Collectives / Speaker
Matthew Lee
AI safety researcher
Matthew works with Poseidon Research and coauthored StegoBench, a study of hidden communication in language models. His New Collectives talk explores collusion and the challenges of monitoring multi-agent systems.
At New Collectives
Talks & conversations.
Understanding Agent Swarms: Multi-Agent Evals · October 21, 2026TALK 04▶ Watch video
▶ Watch videoSelected work
StegoBench
Coauthored research evaluating hidden communication in language models.