graphic showing a black square with AI in white lettering inside with lines terminating in dots coming out of the black square

How Can AI Improve Disaster Planning and Response For Disability-Led Organizations?

Body

The urgency of disaster response intensifies when considering the unique needs of individuals with disabilities. In the chaos of a crisis, locating essential resources—such as wheelchairs, sign language interpreters, direct support professionals, and accessible shelters—becomes a daunting task. Unfortunately, if people with disabilities are excluded from disaster preparedness, their vital needs may be overlooked, leaving them without the equipment or assistance they require to stay safe.

The (WID) has been pioneering efforts to connect disability-led organizations with critical resources to ensure better preparation, response, and recovery during disasters through its (GADRA) program. GADRA works globally to unite disability-led organizations, corporations, and government agencies to expedite the identification and delivery of vital resources to areas impacted by disasters. Since its inception in 2020, GADRA, along with its partners, has responded to multiple humanitarian crises, including the refugee crisis in Gaza, the ongoing war in Ukraine, and natural disasters like the Los Angeles wildfires.

However, the process of locating specialized resources during a disaster — particularly equipment and support tailored to the needs of people with disabilities — remains both time-consuming and complex. , Chief Program Officer at WID and one of the leaders of GADRA, explains that the search for the right resources often becomes a bottleneck in disaster relief efforts. To streamline this process, GADRA is collaborating with engineers offering pro bono services to develop an AI-powered tool that will facilitate faster connections between organizations and the resources they need.

"The NGOs we partner with are disability-led organizations that are either actively responding to a disaster or proactively preparing for one. Initially, we ask them to complete an application that helps us understand their needs—and, more importantly, the needs of the communities they serve," Skaggs explains. "One challenge we’ve noticed is that, during disasters, it’s incredibly difficult for organizations to quickly and systematically assess their priorities. The AI tool will provide a structured logic that guides organizations through a process to clearly identify their most urgent needs and make decisions faster, whether it's for immediate or future requirements."

The first aspect of the AI tool is designed to help organizations prioritize needs more efficiently. The second piece, a ‘web scraping’ tool, will mine data from websites, reports, and other online sources to identify where resources—such as durable medical equipment, prosthetics, or ASL interpreters—might be located.

“The web scraping tool will pull data not just from websites, but also from important documents like after-action reports and tax filings that may contain valuable insights about resource availability,” says Skaggs. “For example, the AI might uncover that a large humanitarian aid organization delivered 25 wheelchairs to Uganda after a previous flood and plans to stage another 25 due to the recurring nature of the disaster. If a flood occurs in Tanzania, the tool could quickly identify that similar resources may be available, and they could be mobilized in time to assist those in need.”

Although there will still be challenges in finding certain resources, the web-scraping feature will enable GADRA to approach disaster response with a clearer understanding of resource availability, helping to prioritize efforts globally. “If it functions as we envision, this tool could turn months of searching into mere days,” Skaggs adds.

Additionally, the AI tool will match organizations with available resources, taking into account shared geographic locations and communities served. This will enhance collaboration and improve the efficiency of resource allocation.

GADRA plans to share information gathered by the AI tool with disability-led organizations, allowing them to pursue their own outreach and resource sourcing if they prefer. "The engineering companies we’ve partnered with have been exceptional, and what makes this project even more impactful is that some of the engineers are people with lived experience of disability. This ensures that disability leadership is embedded throughout the process," says Skaggs. “We are also taking great care to ensure the AI system operates without ‘hallucinations,’ bias, or errors, and that it provides accurate, reliable data.”

In January, GADRA received a prototype of the AI tool and anticipates the full version will be operational by this fall. 

Allen Heinemann, PhD, Director, Center for Rehabilitation Outcomes Research at Shirley Ryan 小恩雅 is a member of the Spinal Cord Injury Model Systems special interest group on disaster preparedness. The group works to ensure that that people with spinal cord injury know how to be prepared for a wide variety of emergencies, including climate events, earthquakes and other natural disasters. The group is working on a fact sheet about preparedness, including considering the status of home accessibility without electricity.

“While the special interest group isn’t currently considering AI as it relates to disaster planning and recovery, it is certainly something we are eager to explore based on WID’s experiences,” Heinemann says.