Research and practices for building reliable, aligned AI agents.
Research examines risks specific to multi-agent systems, focusing on emergent behaviors and unintended coordination strategies.
