Project Abstract

PM2.5 pollution, caused by a complex interaction of agricultural practices, industrial activity, weather conditions, and regulatory enforcement, poses significant health and environmental challenges in many countries, including Thailand and Japan. This project aims to develop an Agentic AI system that autonomously integrates, analyzes, and reasons over multi-source data to provide location-specific recommendations for PM2.5 mitigation and pollution control strategies. The system will utilize AI frameworks (e.g., multi-agent planning, self-directed reasoning, and autonomous task execution) to ingest data from satellite imagery, weather reports, forest fire incident databases, air quality sensors, industrial reports, and local regulations. By comparing scenarios in Thailand and Japan, the model will explore how socio-environmental factors and government policy shape pollution sources and responses. The goal is to generate explainable, data-driven action plans—such as alerts for burn bans, control strategies for industrial emissions, or AI-guided control of air filtration systems—tailored to geographic and regulatory contexts. This project not only advances the capabilities of Agentic AI in environmental decision-making but also contributes to the development of practical tools for public health and urban sustainability.