Skip to content

Topic

Multi-agent AI safety

The study of how several AI agents behave when they can observe, signal to or influence one another, including collusion, unintended cooperation and information leakage between separately run agents.

Current clusters