AgentSociety 2: An Integrated Research Environment for Executable Social Science

2026-06-11

AI Summary

This work addresses a critical gap in computational social science: the absence of executable environments that tightly couple research workflows with simulated societies. Existing systems often treat agents as passive experimental subjects or support only isolated tasks. To overcome this limitation, the authors propose an integrated research platform featuring a novel dual-role large language model agent architecture, wherein agents simultaneously act as AI social scientists and silicon-based participants. This design enables an end-to-end workflow—from hypothesis generation and simulation to result interpretation and manuscript drafting. By integrating configurable social simulations, automated research orchestration, and human-in-the-loop control mechanisms, the platform directly translates theoretical assumptions into auditable agent behaviors and intervention strategies. Across seven cross-scale case studies, the system successfully replicates established qualitative patterns, uncovers meaningful deviations, and demonstrates robust applicability and scalability in micro-experimental, social media, and urban settings.

Abstract

AI scientist systems are beginning to automate parts of scientific research, but social science poses a distinct challenge: its objects of inquiry are not merely datasets or laboratory protocols, but integrated social processes involving situated participants, interaction contexts, interventions, and outcomes. Yet a critical link is missing: existing systems either assist isolated research tasks or simulate agents as experimental subjects, leaving the research workflow and simulated society decoupled. Here we introduce AgentSociety 2, an Integrated Research Environment for executable social science. It couples two roles of LLM agents in the same runtime: AI social scientists that coordinate literature grounding, hypothesis generation, experiment design, simulation execution, result interpretation, and manuscript drafting; and silicon participants that generate behavioral responses within configurable social environments. This dual-role design turns hypotheses into auditable agent behaviors, environment rules, interventions, and measurements, thereby supporting an end-to-end workflow. Across seven illustrative studies spanning micro-level social-science laboratory experiments, meso-level dynamics in social media, and macro-level urban scenarios, we demonstrate its capacity to support diverse disciplinary questions, reproduce major qualitative patterns from prior studies, identify informative deviations, and enable large-scale simulations through optimized agent-environment interactions. By preserving human researchers'high-level agency while delegating procedural orchestration to agentic systems, it provides a human-in-the-loop and controllable infrastructure for next-generation computational social science, with broader applications in scalable computational social experimentation and AI-enabled social governance platforms.

Problem

• computational social science
• agent-based simulation
• AI scientist
• executable social science
• integrated research environment

Innovation

• executable social science
• dual-role LLM agents
• integrated research environment
• silicon participants
• human-in-the-loop simulation