MACC: Multi-Agent Collaborative Competition for Scientific Exploration

Researchers introduce MACC (Multi-Agent Collaborative Competition), a new institutional architecture that combines multiple AI agents based on large language models to improve scientific discovery. The system addresses limitations of single-agent approaches by incorporating incentive mechanisms, shared workspaces, and institutional design principles to enhance transparency, reproducibility, and exploration efficiency in scientific research.

• →Scientific discovery currently relies too heavily on manual individual research efforts, leading to limited exploration and reduced reproducibility.
• →MACC introduces a blackboard-style shared scientific workspace where multiple LLM-based agents can collaborate and compete.
• →The system incorporates institutional mechanisms like incentives and information sharing to encourage better scientific practices.
• →Most existing multi-agent science studies assume single organizational control, limiting examination of how institutional factors affect collective exploration.
• →MACC serves as a testbed for studying how institutional design influences scalable and reliable multi-agent scientific exploration.