New UH project combats food insecurity through AI

One in eight Texans is food insecure, according to the nonprofit Feeding America. That means 1.4 million Texas households are food insecure, with limited or inconsistent access to nutritious food for active and healthy lives. The USDA’s most recent survey on the matter reports that Texas is among the top nine US states with a higher prevalence of food insecurity than the national average.

To solve this problem, a team led by the University of Houston is developing an artificial intelligence-based platform that can support a food charity ecosystem through data-driven technology.

“Our team’s commitment is to helping our neighbours,” said Ioannis Kakadiaris, principal investigator and Hugh Roy and Lillie Cranz Cullen Distinguished University Computer Science Professor in the UH College of Natural Sciences and Mathematics. “This is evident in everything we do and permeates all of our work.”

Funded by a $750,000 grant from the National Science Foundation, this project aims to help food pantries communicate with other pantries, food donors, and agencies while also helping provide clients with culturally conscious and personalized meals.

From the demand side, there must be identification of nutritional needs, cultural preferences, and food preparation equipment as well as supplies of food insecure households, according to Kakadiaris. If one does not know what a particular food is or how to prepare it, it will be wasted, and the efforts of the food charity ecosystem will fail, he added. On the supply side, it is necessary to simplify logistics, improve communication, and coordinate efforts between various stakeholders in the food alms system to optimize the supply chain so that inefficiencies can be minimized.

The platform could potentially use food delivery services such as DoorDash to transfer food. In turn, food donors may be rewarded for their charitable contributions.

“Donors can receive NFTs (non-fungible tokens) which show how well the donor has done in the past month,” said Kakadiaris. “I envision having a gold, silver or bronze donor depending on how much food they have donated over the past month or week.”

The research team from UH included Norma Olvera, professor of education and USDA E. Kika de la Garza Fellow; Elizabeth Anderson-Fletcher, professor of supply chain management at CT Bauer College of Business and Hobby School of Public Affairs; and Susie Gronseth, professor of education. From the University of Texas is Junfeng Jiao, professor and director of the Urban Information Lab in the School of Architecture.

“We will offer users and stakeholders healthy and culturally appropriate recipes using this platform,” said Olvera.

Jiao added that they will ensure AI is fair, secure, transparent and accessible to all.

“This is a multidisciplinary team that brings a variety of expertise to the table,” said Anderson-Fletcher. The team is partnered with Alison Reese, executive director of the Souper Bowl of Caring. Souper Bowl of Caring, home of Tackle Hunger Map, is a non-profit organization that uses its digital platform to raise money for food charities small and large across the country.

This UH project is one of sixteen projects awarded nationwide, totaling $11 million through the NSF Convergence Accelerator program focused on advancing regenerative farming practices and providing nutritious, equitable and affordable food options.

“Food and nutrition security is a new focus for the Convergence Accelerator portfolio, and we are excited to welcome these teams to our program,” said Douglas Maughan, head of the NSF Convergence Accelerator. “We hope to create a synergistic group of efforts that advance regenerative farming practices, reduce water use, provide equitable access to nutritious and affordable food for disadvantaged communities, and spur technology and job creation.”

The Kakadiaris team has been funded through Phase 1 of their project. The Convergence Accelerator Team will submit a formal Phase 2 proposal for up to $5 million in additional support.