Flood Resilience Education through GPT-Guided Gaming with Multi-Scale Built Environment Data
Research Team Members
Approach:
The project addresses the challenge of educating socially vulnerable youth about flood risks by developing an AI-powered educational game that blends real-world data with interactive storytelling. Using the game engine, the team will build immersive virtual environments based on flood-prone communities, integrating socio-demographic data and realistic flood simulations from sources like the National Water Model. Artificial Intelligence will power dynamic, responsive dialogues and role-playing scenarios, turning flood education into an engaging, personalized experience. Educational content: case studies, articles, and crisis communication lessons, will be embedded to support informed decision-making. The game’s impact will be evaluated through user testing, surveys, and analytics (e.g., playtime, quiz performance), with feedback gathered via Amazon Mechanical Turk and community engagement platforms. A beta version will be released in Year 1, with full deployment in Year 2. The strategy leverages interdisciplinary expertise and cutting-edge tech to foster flood resilience and inspire future NOAA careers.
Impact:
1) We will reduce barriers to flood education by making complex information engaging and easy to understand. 2) We will inspire future NOAA professionals by introducing students to climate science and engineering fields. 3) We will generate scalable, open-source educational tools for widespread use in schools, communities, and public agencies. 4) We will promote equity in resilience planning by customizing content to underserved areas using multi-scale built environment data.Abstract:
This project creates an AI-powered educational video game that teaches young people, especially in socially vulnerable communities, about flood risk and resilience. It uses AI-driven storytelling, interactive role-play, and real flood data to simulate flood scenarios in a virtual environment modeled on actual communities. As more frequent and severe flooding, many youth remain uninformed due to unengaging or overly technical flood education tools. Traditional methods are not designed for younger, diverse audiences and often exclude realistic environmental data. To address this gap, we will integrate gamification, built environment modeling, and AI-generated content to create a highly interactive and tailored flood resilience education experience.
In this project, a publicly available, AI-powered educational gaming platform for flood resilience training will be developed. As flood education is largely passive, static, or text-based, existing gamified tools lack realism, interactivity, and relevance to users' lived experiences. The project contributions include:
Advance 1: Gamified Flood Education
• Develop a game with realistic 3D environments using game engine.
• Integrate real-time flood simulation data (e.g., NWM, FIM) for localized scenarios.
Advance 2: AI-Driven Learning
• Use AI techniques to power interactive, scenario-based dialogue with characters.
• Embed crisis communication and risk awareness lessons into roleplay narratives.
Advance 3: Impact Assessment Tools
• Evaluate user engagement through Amazon MTurk and in-game analytics.
• Continuously refine the platform using user feedback, quizzes, and retention tests
Additionally, this project will have the specific impacts as below:
1) We will reduce barriers to flood education by making complex information engaging and easy to understand.
2) We will inspire future NOAA professionals by introducing students to climate science and engineering fields.
3) We will generate scalable, open-source educational tools for widespread use in schools, communities, and public agencies.
4) We will promote equity in resilience planning by customizing content to underserved areas using multi-scale built environment data.