When Agents Evolve, Institutions Follow

2026-04-30

AI Summary

This study addresses the challenge of achieving efficient coordination in multi-agent systems under conditions of limited individual cognition and incomplete information. It presents the first systematic mapping of seven historical human political institutions into multi-agent governance architectures, evaluating their collective intelligence within a unified framework through cross–large language model and cross-task benchmarking. The findings reveal that governance topology exerts a decisive influence on system performance: under the same model, the performance gap between the best and worst institutional designs exceeds 57 percentage points. Moreover, the optimal governance architecture systematically varies with both model capabilities and task characteristics, thereby advancing multi-agent systems toward a self-evolving governance paradigm.

Abstract

Across millennia, complex societies have faced the same coordination problem of how to organize collective action among cognitively bounded and informationally incomplete individuals. Different civilizations developed different political institutions to answer the same basic questions of who proposes, who reviews, who executes, and how errors are corrected. We argue that multi-agent systems built on large language models face the same challenge. Their central problem is not only individual intelligence, but collective organization. Historical institutions therefore provide a structured design space for multi-agent architectures, making key trade-offs between efficiency and error correction, centralization and distribution, and specialization and redundancy empirically testable. We translate seven historical political institutions, spanning four canonical governance patterns, into executable multi-agent architectures and evaluate them under identical conditions across three large language models and two benchmarks. We find that governance topology strongly shapes collective performance. Within a single model, the gap between the best and worst institution exceeds 57 percentage points, while the optimal architecture shifts systematically with model capability and task characteristics. These results suggest that collective intelligence will not advance through a single optimal organizational form, but through governance mechanisms that can be reselected and reconfigured as tasks and capabilities evolve. More broadly, this points to a transition from \textbf{self-evolving agents} to the \textbf{self-evolving multi-agent system}. The code is available on \href{https://github.com/cf3i/SocialSystemArena}{GitHub}.

Problem

• collective action
• coordination problem
• multi-agent systems
• governance
• institutional design

Innovation

• multi-agent systems
• governance architecture
• collective intelligence
• institutional design
• large language models