Pioneering research at Texas A&M University is leveraging the speed and computing power of artificial intelligence to revolutionize disaster management, aiming to save lives and significantly reduce the impact of hazardous weather events in Texas and across the nation.
Dr. Ali Mostafavi a professor of civil and environmental engineering views Texas a frequent site of hurricanes, flash floods, wildfires, and tornadoes as “ground zero for natural disasters” and Texas A&M as “ground zero for solutions.” He highlights that expediting the prediction of a flooded neighborhood by even 30 minutes can save hundreds of lives.
Researchers at his UrbanResilience.AI Lab, in a new partnership with Meta are developing AI systems designed to augment situational awareness and resilience across every stage of a weather hazard.
The AI Advantage: Speed and Insight
The core advantage of these AI systems is their ability to rapidly analyze massive, disparate datasets that are too complex for humans to process quickly during a crisis. For an upcoming flash flood, for example, AI can analyze readings from rainfall sensors and stream gauges alongside 20 years of historical flooding data to instantly predict the neighborhoods and ZIP codes most likely to be impacted.
“AI systems have significant computing power and speed to reliably provide the insights humans need to make decisions during times of crisis when every minute counts,” said Dr. Mostafavi.
As part of the Meta partnership, the lab will utilize large language models, such as Llama, to create Disaster Management Companion AI systems. Dr. Mostafavi stated, “If we are able to integrate these applications into disaster management processes, we can significantly improve the anticipation, situational awareness and response to these events.” He added that as these events become more frequent, AI systems will serve as the “next-defense barriers for cities and communities.”
Solutions Across the Disaster Cycle
Dr. Mostafavi’s lab is at the forefront of using predictive, analytic, and generative AI to manage the entire disaster cycle:
• Mitigation and Preparedness: Predictive models can anticipate how people will evacuate from high-impact areas and evaluate which neighborhoods face the highest flood risk before a storm makes landfall.
• Response: During an active crisis, AI monitors near real-time data on neighborhood-level evacuations, power outages, and property damages, guiding decision-makers on where to deploy search and rescue personnel and other critical resources.
• Recovery: AI can perform rapid impact assessment using high-resolution satellite and street-level imagery.
These technologies are not meant to replace humans but to augment human decision-making. The lab is also developing an emergency operating center “companion” that functions like a custom-built ChatGPT for emergency managers, providing instant access to historical reports and data sources.
Dr. Mostafavi and his team have field-tested these tools during real-world events, including Hurricanes Beryl, Milton, and Helene, and the Los Angeles wildfires, a process essential for refining the technology. He anticipates that multiple AI applications will become standard tools for emergency management within the next three to five years, stressing the need for continued funding and field-testing to ensure they can be scaled and deployed nationwide.