City emergency managers want to employ artificial intelligence to help detect natural hazards across Oʻahu.
The Honolulu City Council passed a 3 last year urging city officials to install monitoring devices in areas vulnerable to flooding. After the Kona lows this spring, lawmakers then allocated $300,000 to the city's Department of Emergency Management to support the development of a flood detection system.
DEM Director Randal Collins provided an update on that work to the council on Tuesday. He told lawmakers he was confident that a pilot of the early detection system could be up and running within this fiscal year, which ends in June.
Collins said DEM plans to leverage monitoring systems that are already in place on the island, as well as fill gaps in existing hazard surveillance with cameras and gauges.
The department plans to use machine learning to help analyze data from the monitors and flag emerging threats for emergency officials. While Collins said that AI had an important role to play in hazard monitoring, he also emphasized that an emergency official would review and verify all the AI-powered system's conclusions.
“There will always, always be a human element to this,” Collins said.
DEM plans to convene a working group with “basically anybody that has a camera system that we're able to tap into,” according to Collins, including the National Weather Service, the United States Geological Survey, the University of Hawaiʻi and AI-developers EM1 and Hohonu.
Those stakeholders will help DEM determine what infrastructure to buy and how to integrate it into DEM's operations. Assessing flooding hazards will be the pilot's first focus, but Collins also noted that Honolulu would benefit from robust early detection systems for threats such as wildfire, tsunami, and even cyber attacks.
