Cooperate to Compete

Strategic Coordination in Multi-Agent Conquest

Most current multi-agent benchmarks test pure cooperation, pure competition, or structured bargaining. Cooperate to Compete (C2C) captures a more realistic interaction paradigm: agents with misaligned interests use negotiation as a tool to achieve individual goals. We design a mixed-motive environment where humans and AI must navigate both short-term cooperation and long-term competition. By studying both agents and humans in this environment, we analyze their behavioral differences and introduce interventions to evaluate how specific negotiation behaviors impact performance.


Contributors

Abigail O’Neill, Alan Zhu, Mihran Miroyan, Narges Norouzi, Joseph E. Gonzalez