Text to end hunger: AI system coordinates food aid in emergencies

Harnessing AI, US researchers at the University of Houston (UH) are developing a new online resource for national food pantries to expedite food aid and streamline stakeholder collaboration in the aftermath of natural disasters.

Food pantry leaders working with this platform could send SMS text messages during an emergency, which the AI system would process and organize in order of priority on its dashboard.

The UH researchers are directing US$300,000 of funding from the US Department of Agriculture (USDA) toward publishing a website featuring the tool for Florida food pantries, which could reduce communication bottlenecks and later support regions like Houston.

“Our AI system can automatically and efficiently process large volumes of status reports, enabling emergency coordinators to respond more quickly when demand spikes,” says Ioannis Kakadiaris, principal investigator and professor of computer science in UH’s College of Natural Sciences and Mathematics.

The system will also be adaptable to various disaster scenarios, with typical support including food, water, first aid, and other emergency supplies.

“We want the tool to be completely flexible, so if there are fires or some other type of damage or disaster, it can handle the new situation,” comments Marcus Sammer, application developer for the project at Kakadiaris’ Computational Biomedicine Lab.  

Learning from Hurricane Harvey

The research team is building on experience gained during Hurricane Harvey in 2017. Kakadiaris adds: “We expect the lessons learned in Florida will strengthen our disaster response capabilities back home in Houston.”

Disasters often worsen food insecurity, stresses the research team. Access to essential food resources can be blocked by power outages, floods, wildfires, or tornadoes.

Meanwhile, USDA reported a 12.8% rise in food insecurity from 2022 to 2023, when 13.5% of US households were food insecure at least one time.

In June, the USDA’s National Institute of Food and Agriculture awarded UH and three other institutions nearly US$1.2 million to tackle disaster response efforts after hurricanes Helene and Milton devastating Florida last fall.

The project will also build on two previous grants Kakadiaris received from the US National Science Foundation. The researcher has received more than US$2.2 million since 2021 to scale AI-based food security systems.

Over the next year, the UH research team will survey stakeholders to better understand current communication systems. They plan to release a pilot version of the system to Florida food pantries for testing and feedback by September.

“Our job is to explore what the challenges are that need to be addressed using the technology, and then we hope that we or somebody else will run with this technology,” says Kakadiaris.

In a similar use case for text messaging to facilitate hunger relief, UNICEF’s SMS-based system, RapidPro, rolled out earlier this year to streamline the management of nutritional stock supplies at Chad health centers. The system enables health center managers to receive delivery alerts, report stock levels, and trigger automatic alerts for potential shortages. The organization reports the tool is scalable, simple to use, and free of charge.