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

Name: High Volatility and Action Bias Distinguish LLMs from Humans in Group Coordination
Uploaded: 2026-10-03T03:15:16.406648+00:00
Duration: 517 s
Description: This study compared human and Large Language Model (LLM) performance in a common-interest game called Group Binary Search (GBS). Humans improved their coordination over multiple games, while LLMs showed minimal learning and excessive switching. LLMs overreacted to feedback, maintained high volatility, and failed to stabilize group behavior.

This study compared human and Large Language Model (LLM) performance in a common-interest game called Group Binary Search (GBS). Humans improved their coordination over multiple games, while LLMs showed minimal learning and excessive switching. LLMs overreacted to feedback, maintained high volatility, and failed to stabilize group behavior.

Open the full episode