High Volatility and Action Bias Distinguish LLMs from Humans in Group Coordination

Research comparing large language models (LLMs) to humans in group coordination tasks reveals that LLMs exhibit excessive volatility and switching behavior that impairs collective performance. Unlike humans who adapt and stabilize over time, LLMs fail to improve across repeated coordination games and don't benefit from richer feedback mechanisms.

• →LLMs demonstrate poor group coordination abilities compared to humans in common-interest games requiring collective action.
• →LLMs exhibit excessive switching behavior and high volatility that prevents groups from converging on solutions.
• →Humans adapt and stabilize their coordination strategies over time while LLMs fail to show learning across repeated games.
• →Richer feedback significantly benefits human coordination but has minimal impact on LLM performance.
• →The research identifies key behavioral differences that highlight current limitations in LLM collective intelligence capabilities.