LLM-Based Chatbot for Flood Assistance

• Eduardo A. Reolon UTFPR
• Ricardo Lüders UTFPR
• Luiz Gomes-Jr. UTFPR

DOI: https://doi.org/10.5753/erbd.2025.6862

Abstract

Flood events pose a significant risk to the population, especially in vulnerable urban areas. Studies indicate that many fatal incidents occur due to risky behaviors adopted by the victims, making the dissemination of accurate information essential during such situations. This work presents an intelligent agent based on Large Language Models (LLMs), designed to provide real-time guidance to people affected by floods. The agent processes contextual information, such as weather forecasts, geospatial data, and urban infrastructure, to generate personalized recommendations on safe movement, staying at home, and seeking shelter. The evaluation of the agent, conducted through a questionnaire with different emergency scenarios, demonstrated a positive performance in communication and assistance to users, highlighting the potential of artificial intelligence for the mitigation of natural disasters.

References

Basnyat, B., Roy, N., and Gangopadhyay, A. (2020). Towards AI conversing: Floodbot using deep learning model stacks. In Proceedings of the IEEE/ACM International Conference on Cyber-Physical Systems (ICCPS). IEEE.

Davis, F. D. (1993). User acceptance of information technology: system characteristics, user perceptions and behavioral impacts. International Journal of Man-Machine Studies, 38(3):475–487.

Diakakis, M. (2020). Types of behavior of flood victims around floodwaters. Correlation with situational and demographic factors. Sustainability, 12(11):4409.

Diaz-De-Arcaya, J., López-De-Armentia, J., Miñón, R., Ojanguren, I. L., and Torre-Bastida, A. I. (2024). Large language model operations (LLMOps): Definition, challenges, and lifecycle management. In 2024 9th International Conference on Smart and Sustainable Technologies (SpliTech), pages 1–7. IEEE.

Fernandez, H. G. and Splendore, P. R. (2021). Sistema de identificação automática de riscos hidrometeorológicos com retroalimentação e reestruturação autônoma da infraestrutura de comunicação. Master’s thesis, Universidade Tecnológica Federal do Paraná.

Kasinathan, V., Mustapha, A., Anuar, A. S. Z., and Lim, E. C. M. (2020). Flow: Intelligent flood aid chatbot as virtual flood assistant. International Journal of Management, 11(10):1749–1756.

LangChain (2025). Langchain: Building applications with LLMs. Acessado em: 31 mar. 2025.

Ollama (2023). Meta’s llama 3.2 goes small with 1b and 3b models. Disponível em: [link].

Urbanelli, A., Frisiello, A., Bruno, L., and Rossi, C. (2024). The ERMES chatbot: A conversational communication tool for improved emergency management and disaster risk reduction. International Journal of Disaster Risk Reduction, 112:104792. Under a Creative Commons license.

von der Mosel, J., Trautsch, A., and Herbold, S. (2023). On the validity of pre-trained transformers for natural language processing in the software engineering domain. IEEE Transactions on Software Engineering, 49(4):1487–1507.