Problem
Research questions and friction points this paper is trying to address.
social science automation
human participants
experimental scalability
labor-intensive research
methodological complexity
LLM Agents as Social Scientists: A Human-AI Collaborative Platform for Social Science Automation
🤖 AI Summary
This study addresses the limitations of traditional social science research—such as high costs and poor scalability due to reliance on human participants—by introducing S-Researcher, a novel platform that “siliconizes” the entire research workflow through large language model (LLM) agents. Integrating inductive, deductive, and abductive reasoning, S-Researcher establishes a scalable, generalizable paradigm for human-AI collaborative inquiry. Built upon the YuLan-OneSim large-scale simulation system, the platform enables natural language–based automatic programming, supports hundreds of thousands of concurrent agents, and incorporates feedback-driven LLM fine-tuning. Empirical evaluations successfully replicate cultural dynamics, validate the teacher attention hypothesis, and uncover cooperation mechanisms in public goods games, with findings corroborated through human-subject experiments.
📝 Abstract
Traditional social science research often requires designing complex experiments across vast methodological spaces and depends on real human participants, making it labor-intensive, costly, and difficult to scale. Here we present S-Researcher, an LLM-agent-based platform that assists researchers in conducting social science research more efficiently and at greater scale by "siliconizing" both the research process and the participant pool. To build S-Researcher, we first develop YuLan-OneSim, a large-scale social simulation system designed around three core requirements: generality via auto-programming from natural language to executable scenarios, scalability via a distributed architecture supporting up to 100,000 concurrent agents, and reliability via feedback-driven LLM fine-tuning. Leveraging this system, S-Researcher supports researchers in designing social experiments, simulating human behavior with LLM agents, analyzing results, and generating reports, forming a complete human-AI collaborative research loop in which researchers retain oversight and intervention at every stage. We operationalize LLM simulation research paradigms into three canonical reasoning modes (induction, deduction, and abduction) and validate S-Researcher through systematic case studies: inductive reproduction of cultural dynamics consistent with Axelrod's theory, deductive testing of competing hypotheses on teacher attention validated against survey data, and abductive identification of a cooperation mechanism in public goods games confirmed by human experiments. S-Researcher establishes a new human--AI collaborative paradigm for social science, in which computational simulation augments human researchers to accelerate discovery across the full spectrum of social inquiry.
Innovation
Methods, ideas, or system contributions that make the work stand out.
LLM agents
social simulation
human-AI collaboration
auto-programming
scalable agent architecture
🔎 Similar Papers
No similar papers found.
