An Infectious Disease Spread Simulation Based on Large Language Model Decision Making

2026-06-04 · Source: Artificial Intelligence · Field: Science & Research — Artificial Intelligence & Machine Learning, Health & Medical Research, Mathematics & Computational Sciences · Depth: Advanced, quick

Summary

A new spatially grounded, agent-based simulation framework integrates large language model (LLM)-generated decisions to model infectious disease spread. This framework uses a census-based synthetic population, assigning agents to specific spatial units within cities based on real-world census data to capture demographic and geographic distributions. It simulates self-reported influenza-like illness, comparing three decision scenarios: independent reasoning, household influence, and message framing. Simulations conducted in San Francisco and Atlanta revealed that income and education are the primary drivers of variation in reporting rates. Geography, the choice of LLM model, and message framing also showed smaller but consistent effects. The framework generates synthetic data that reflects both social and geographic heterogeneity, supporting spatial epidemiological modeling and bias-aware behavioral analysis.

Key takeaway

For research scientists developing epidemiological models, this framework demonstrates how integrating LLM-driven agent decisions with real-world census data can enhance simulation realism. You should consider incorporating socioeconomic and geographic heterogeneity into your models to better predict disease reporting rates and design more targeted public health interventions. This approach allows for bias-aware behavioral analysis, crucial for effective policy.

Key insights

LLMs can simulate human decision-making in spatially-grounded infectious disease models, revealing key socioeconomic drivers.

Principles

• Socioeconomic factors like income and education strongly influence disease reporting.
• Geographic and social contexts impact individual health behaviors.
• LLMs can model complex human decision-making in simulations.

Method

The framework integrates LLM-generated decisions on self-reported illness into a census-based synthetic population, assigning agents to spatial units using real-world census data to model diverse behaviors and compare decision scenarios.

In practice

• Use LLMs for agent-based modeling of public health behaviors.
• Incorporate census data for spatially diverse simulations.
• Analyze socioeconomic drivers in health intervention design.

Topics

• Infectious Disease Modeling
• Large Language Models
• Agent-Based Simulation
• Public Health Interventions
• Spatial Epidemiology
• Behavioral Modeling

Best for: AI Scientist, Research Scientist

Related on AIssential

• An Infectious Disease Spread Simulation Based on Large Language Model Decision Making — LLMs can simulate human behavior in spatially-grounded disease spread models, revealing demographic drivers.
• An Infectious Disease Spread Simulation Based on Large Language Model Decision Making — LLMs can simulate nuanced human health behaviors in spatially-grounded agent-based models, revealing demographic disparities.
• Neetyabhas: A Framework for Uncertainty-Aware Public Policy Optimization in Rational Agent-Based Models — Public policy optimization benefits from integrating individual behaviors and real-world uncertainties into agent-based models.
• Bayesian inferences and frequentist evaluations — Decision-theoretic sequential experimental design can optimize clinical trial recruitment restarts post-pandemic.
• TRIAGE: Dialectical Reasoning for Explainable Risk Prediction on Irregularly Sampled Medical Time Series with LLMs — TRIAGE uses dialectical LLM reasoning to provide calibrated, continuous risk scores and interpretable rationales for medical time series.
• The Epi-LLM Framework: probing LLM behavioral priors through epidemiological agent-based models — The Epi-LLM framework uses LLMs in agent-based models to simulate epidemic behavior, showing LLMs can reduce infection peaks.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Artificial Intelligence.