San Jose adds AI tools to its disaster tactics

San Jose faces a wide array of disaster risks — everything from wildfires to floods and mass shootings — each of which poses its own difficult questions for emergency management officials tasked with leading the city’s evacuation efforts. But those emergency planners will soon get a high-tech assist in these life-or-death calls from a cutting-edge, artificial intelligence-powered software package.

The City Council signed off on a $3.5 million, six-year agreement with Ladris Technologies to purchase a system its creators said can distill vast quantities of data to determine which route will most quickly lead evacuees to safety. The purchase comes just in time for the technology to aid in emergency preparations for a series of major sports events expected to bring hundreds of thousands of visitors to San Jose next year.

“Trying to evacuate a dense urban area becomes quite a challenge,” Ray Riordan, director of San Jose’s Office of Emergency Management, told San José Spotlight. “There’s so much information you have to try and gather, make assessments and take action in order to preserve public safety.”

For example, to determine how many cars might pour out onto a potential evacuation route, emergency planners might want to look at data on a neighborhood’s population density and rate of car ownership. Emergency officials might also want to assess the route’s safety. To do so, they would likely need to draw on yet more sources of data: flood maps, weather forecasts, traffic conditions and even real-time data from local dispatch centers.

Calculating possible evacuation times from such a dizzying churn of data might take a human several hours. But the Ladris software can cut that down to mere minutes, according to city officials who conducted a nine-month evaluation of the technology.

“In a disaster scenario, one of your biggest resources is time,” Ladris Technologies CEO Leo Zlimen told San José Spotlight. “Time to make the right decisions. Time to physically go and evacuate people. And even before the disaster, time to educate people about how much time they have.”

Ladris, based in Nevada City, California, has already partnered with nearly 100 emergency response agencies across eight states, Zlimen said.

The rollout of the technology fits with San Jose’s broader embrace of AI tools to boost city governance. San José Spotlight previously reported the city is already using AI-powered software to optimize public transit, translate public meetings, review official documents and catalogue street-level blight and safety problems such as potholes, graffiti, broken street lights and illegal dumping.

Meanwhile, emergency planning is underway for the trio of high-profile sports events set to descend on the South Bay next year — Super Bowl 60, NCAA March Madness and FIFA World Cup. Santa Clara County has been holding drills preparing county hospital workers for active shooters, terrorist attacks and natural disasters like earthquakes.

Officials with the Santa Clara County Office of Emergency Management told San José Spotlight that county agencies have not been involved in the rollout of the Ladris software in San Jose. However, Riordan said San Jose will share information produced by the technology for the purpose of coordination and public safety.

He also affirmed the software will not make any evacuation calls independent of the humans who operate it.

“We make the decisions, but it’s a tool that helps us get the information together to make the right decisions,” he said. “You can’t just rely on AI by itself.”

Zlimen said his company has used recent advances in machine learning technology to make predictive modeling software capable of incorporating new, unexpected streams of data.

He pointed to a novel application developed by San Jose officials during test runs of the software. As emergency planners modeled flood scenarios, they wanted to know where homeless residents living next to the city’s creek beds might flee to as the waters rise. The software came up with predictions that could help city agencies more easily locate newly formed homeless encampments and direct resources accordingly, officials said.
It’s an outcome that surprised even Zlimen’s team, he said. Such flexibility will be essential for emergency planners, he argued, as the emergency risks cities face continue to morph and change over the coming decades.

“There’s a huge amount of inputs into those problems, but very few focused metrics coming out,” Zlimen said. “So this was able to (create) a process that was useful for the city, but also, potentially, a great example for how other cities may be able to tackle a similar problem.”

Contact Keith Menconi at [email protected] or @KeithMenconi on X.