A book chapter and a position paper explore the use of Large Language Models (LLMs) in social simulations. The book chapter traces the evolution from agent-based models to AI-enhanced simulations and Social Digital Twins, emphasizing the shift towards more realistic representations of social systems. The position paper argues for the necessity of clear boundaries in LLM-based social simulations, highlighting that current LLMs' tendency towards homogeneous outputs limits their ability to capture essential behavioral diversity. It proposes methods for validation and constraint to ensure these simulations provide genuine insights into social science. AI
IMPACT LLM-based social simulations are advancing, but require careful validation and boundary setting to ensure meaningful insights into social dynamics.
RANK_REASON The cluster consists of academic papers discussing methodologies and theoretical boundaries for using LLMs in social simulations.
- Putnam's Social Capital Theory
- Robert Putnam
- SocaSim
- Large Language Models
- LLM-Based Social Simulations
- norm propagation
- social science
- agent-based model
- arXiv
- Hugging Face
- Social Digital Twins
AI-generated summary · Google Gemini · from 4 sources. How we write summaries →
