Artificial Intelligence and Machine Learning

Cornell Economist Says Disaster Relief Workers Need AI Tools

Cornell President David Skorton and Professor Christopher Barrett
Cornell University President David Skorton, left, with Christopher Barrett, the Stephen B. and Janice G. Ashley Professor of Applied Economics and Management.

 This afternoon we broke out into working groups. I attended the one on interactive decision support tools. Christopher B. Barrett, Professor of Applied Economics and Management International Professor of Agriculture at Cornell, talked about the need for such tools in the face of natural disasters.

In Sri Lanka after the 2004 Tsunami, humanitarian emergency responders were not sure what kind of assistance many survivors needed — cash for food? Food itself? Cash can be inflationary. Should they order food from inland Sri Lanka or wait for wheat to be shipped from North Dakota or from India? If they bought locally, that might reduce costs by 34% and delivery delays by 78%.

These were first-time choices because until then, by law, American humanitarian agencies could only donate American food. “Furthermore, field workers must make decisions in data-sparce environments quickly because lives are on the line,” said Barrett. He showed a photo of a mother feeding her starving child leaves that would squelch his appetite.

Barrett and his colleagues are working with CARE (Cooperative for Assistance and Relief Everywhere), the largest US food-aid agency, to develop tools for response. They have developed a simple decision tree asking two questions to guide agencies on whether to supply food aid or not: Are local markets functioning well? If yes, then give survivors cash.

CARE wanted to implement the system but the sequence of questions to ask varies depending on location.

“We need automatic rules, a smart system for guidance," he said, but “even a course tool is highly conditional and may be too complicated for the users—emergency workers who are generalists, under high stress, and sleep-deprived.

AI could improve efficiency with a system that prompts the next question to ask, gradually building up advice on the best ways to respond. “We need help putting this together and sticking it on a Palm Pilot that we can hand to a Catholic relief worker somewhere so that they can perform assessments and response analysis done in field.”

The number of people affected by such decisions can be 1 million at a time.

Karen A. Frenkel writes about science and technology. She lives in New York City. This is the link to her website.



Join the Discussion (0)

Become a Member or Sign In to Post a Comment

The Latest from CACM

Shape the Future of Computing

ACM encourages its members to take a direct hand in shaping the future of the association. There are more ways than ever to get involved.

Get Involved

Communications of the ACM (CACM) is now a fully Open Access publication.

By opening CACM to the world, we hope to increase engagement among the broader computer science community and encourage non-members to discover the rich resources ACM has to offer.

Learn More