AI technology has the potential to strengthen state and local governments’ natural disaster preparation and response processes, from data modeling to predictions.
Kansas City’s one-day pilot involved leveraging technology to determine whether the city could accelerate the process of filing the Federal Emergency Management Agency (FEMA) preliminary damage assessment — because as the city’s Chief Digital Officer Andrew Ngui said, “every moment counts” following a disaster. Officials piloted this technology March 23.
For Kansas City, the exercise stemmed from the work of Bellwether, a team of technologists at Google’s Moonshot Factory. Bellwether worked with the Defense Innovation Unit — a U.S. Department of Defense organization focused on scaling commercial technology — to identify critical infrastructure that has been damaged following a severe weather event, according to Josh Jeffery, Bellwether senior program manager.
The pilot involved using the city’s drone program to collect aerial images, which were then processed with AI, Jeffery said.
Traditionally, post-disaster damage assessment has been done on paper, Ngui said. It involves city staff and volunteers being trained in disaster management and then deployed to canvas the entire damage zone from the street. Damage might include fallen electrical poles, trees or water — although the pilot was a simulation scenario.
Numerous metrics were used to evaluate the simulation’s impact, Ngui said. First, from a holistic standpoint, being able to send drones as the advance party to evaluate what may be a hazardous location ahead of, or instead of, sending out humans improves safety in the community.
“Safety is No. 1,” said Ngui, citing a goal similar to minimally invasive surgery: “minimally invasive innovation.” After that, it is about accelerating the process of accessing aid. “Because it’s not just about the damage; it’s about really restoring people’s lives as quickly as possible to the status quo.”
The speed with which data processing could be done — about 23 minutes, in this scenario — demonstrated drastic time savings, according to Jeffery. During the pilot, about 60 people were assessing damage from the ground to compare the traditional process with that being tested. Technology allowed for faster data collection and data processing, compared to that done by humans.
It was also important to measure accuracy and understand how well the technology could assess damage using FEMA’s sliding scale from one to five. The technology, Jeffery said, was 96 percent accurate within one number.
The city is about 319 square miles, and when assessing from the street level, some damage may be missed. This technology enables a more comprehensive view of damage around the city — and even beyond city borders.
Kansas City’s emergency management coordinator, Geoff Hinkle, engaged with private and public agency partners for the pilot, Ngui said, including city fire and police, Platte County, and the National Weather Service.
The initiative serves as an example of the potential to transform the emergency management process. What was once a paper-based exercise can now incorporate technology from Esri’s Survey123 tool, a drone system, and AI.
The city has applied for a grant for AI-powered government innovation; if it is selected, officials will focus on making infrastructure more resilient against disasters.
For the pilot, the focus was on one Kansas City neighborhood, but Jeffery’s vision for the technology is broader: that residents of any city could benefit from increased knowledge of the impacts a disaster might have on one’s home.
This exercise could help create a playbook for rapid post-disaster damage assessment for other cities, Ngui said. Officials plan to engage with stakeholders in the city and elsewhere.
The pilot is informing Bellwether’s development process, too, Jeffery said, so that the team could offer these technological capabilities to municipalities across the country, even those without a drone program like that of Kansas City. Beyond disaster response, he said, a key goal is to inform disaster planning so cities are better prepared against future events.
