Using artificial intelligence when disaster strikes

Artificial intelligence could play a critical role in all phases of disaster resilience, according to “Into the Storm: Using Artificial Intelligence to Improve California’s Disaster Resilience,” an issue brief the Partnership released with Microsoft in early July. The brief explores the AI tools governments can use for disaster resilience and highlights how agencies are using AI technologies in the field.

Considering that the Federal Emergency Management Agency has declared 27 major disasters across the United States in 2020 so far—not including those related to the COVID-19 pandemic—government officials need to think through how they can strengthen their ability to prepare for, respond to and recover from these shocks. Disasters put lives and livelihoods at risk, and governments at all levels—federal, state and local—must seek the best tools available to tackle the complex challenges involved and protect lives.

At a release event held in July, Bijan Karimi, assistant deputy director for emergency services at the San Francisco Department of Emergency Management, and Stuart McKee, chief technology officer for state and local government at Microsoft, discussed basic principles disaster resilience officials should keep in mind when considering the use of AI.

• Play to AI’s strengths. AI can quickly collect and analyze large amounts of data and information. Karimi noted how closely that capability aligns with his team’s work of gathering, analyzing and disseminating information during disasters. But protracted information-gathering can slow data analysis and delay dissemination. “We spend an inordinate amount of time just gathering information, and I can’t get my [team] enough time to focus on the abstract reasoning,” Karimi said. AI tools could be used to automate the process of gathering and organizing data, leaving more time for disaster resilience professionals and first responders to act on the information.
• AI can only do so much. AI is a helpful tool but not a replacement for smart, experienced emergency managers and first responders. Information from AI tools that predicts the path of a wildfire or identifies buildings that have sustained extensive damage after an earthquake isn’t helpful without disaster resilience professionals to interpret and act on the information and analyses. “Technology, at the end of the day, doesn’t solve problems,” McKee said. “It takes smart people using technology creatively [to come up with solutions].”
• Use robust and trustworthy data. AI tools are only as reliable as the data they use, so good data is the cornerstone of any successful AI tool. When it comes to the potential destruction of disasters, governments must use high-quality, accurate data that also incorporates information about the lives, livelihoods and communities that may be particularly vulnerable to a disaster’s impact.

Read the report to learn more about potential uses for AI tools in disaster preparation, response and recovery, as well as additional considerations for governments seeking to implement AI for disaster resilience.

For more on how the government can solve our nation’s biggest challenges, read:

• Learning from the government’s successful use of emerging technologies
• Risky business: A framework for risk management in federal government
• Leaders show how government can improve hiring to address 21st century tech challenges

Elizabeth Byers supports the Partnership’s research on government effectiveness, including projects around the use of data and technology in government.