Modeling human reputation-seeking behavior in a spatio-temporally complex public good provision game

2025-06-06

AI Summary

This study investigates reputation-driven cooperation mechanisms for public goods provision in spatiotemporal complex environments. We propose a multi-agent reinforcement learning (MARL)-based reputation modeling framework and validate it—within the Clean Up spatiotemporal social dilemma—using real human behavioral data, marking the first such empirical validation of reputation-guided cooperation in MARL. Results show that under identifiable-identity conditions, both humans and agents spontaneously evolve turn-taking strategies, achieving high cooperation success rates (>85%); under anonymity, cooperation rapidly collapses (<20%). This emergent turn-taking mechanism transcends classical game-theoretic predictions, revealing that identity identifiability sustains cooperation not primarily through punishment, but by enabling strategic reciprocity. Our findings provide critical theoretical foundations for modeling socially situated agents and designing institutional mechanisms that leverage reputational incentives.

Technology Category

• Multiagent Systems: Mechanism Design
• Game Theory and Economic Paradigms: Cooperative Game Theory
• Humans and AI: Learning Human Values and Preferences

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Economics, Online Markets and Human Computation: Social networks and social learning
• User Modeling, Personalization and Recommendation: Accountability, Transparency, and Ethics for personalization

Abstract

Multi-agent reinforcement learning algorithms are useful for simulating social behavior in settings that are too complex for other theoretical approaches like game theory. However, they have not yet been empirically supported by laboratory experiments with real human participants. In this work we demonstrate how multi-agent reinforcement learning can model group behavior in a spatially and temporally complex public good provision game called Clean Up. We show that human groups succeed in Clean Up when they can see who is who and track reputations over time but fail under conditions of anonymity. A new multi-agent reinforcement learning model of reputation-based cooperation demonstrates the same difference between identifiable and anonymous conditions. Furthermore, both human groups and artificial agent groups solve the problem via turn-taking despite other options being available. Our results highlight the benefits of using multi-agent reinforcement learning to model human social behavior in complex environments.

Problem

• Modeling human reputation-seeking in complex public good games
• Validating multi-agent reinforcement learning with human experiments
• Comparing cooperation strategies in identifiable vs anonymous conditions

Innovation

• Multi-agent reinforcement learning models reputation-based cooperation
• Identifiable vs anonymous conditions affect cooperation success
• Turn-taking strategy solves complex public good provision