Pimmur, can LLM simulate human collective behavior?
The article discusses the PIMMUR principles aimed at ensuring validity in the simulation of human collective behavior by large language models (LLMs). A systematic audit of 39 studies revealed that a significant majority violate at least one of these principles, raising concerns about the reliability of AI simulations. The findings suggest that many observed behaviors in LLM societies may be artifacts of methodology rather than true representations of human social dynamics.
- ▪Large language models are increasingly used to simulate human collective behaviors.
- ▪An audit of 39 studies found that 89.7% violated at least one of the PIMMUR principles.
- ▪Many reported collective phenomena disappear or reverse when PIMMUR principles are enforced.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Computation and Language arXiv:2509.18052 (cs) [Submitted on 22 Sep 2025 (v1), last revised 6 Apr 2026 (this version, v3)] Title:The PIMMUR Principles: Ensuring Validity in Collective Behavior of LLM Societies Authors:Jiaxu Zhou, Jen-tse Huang, Xuhui Zhou, Man Ho Lam, Xintao Wang, Hao Zhu, Wenxuan Wang, Maarten Sap View a PDF of the paper titled The PIMMUR Principles: Ensuring Validity in Collective Behavior of LLM Societies, by Jiaxu Zhou and 7 other authors View PDF HTML (experimental) Abstract:Large language models (LLMs) are increasingly deployed to simulate human collective behaviors, yet the methodological rigor of these "AI societies" remains under-explored.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv.org.
Discussion
0 commentsMore from arXiv.org
