When conflict disrupts agriculture, humanitarian organizations often face the critical challenge of making decisions with limited information about what is happening on the ground.
A new initiative at Arizona State University seeks to provide earlier, more accessible insights into emerging food security risks by combining artificial intelligence, or AI, satellite observations and natural-language tools.
The project, led by Ana M. Tárano, an assistant research professor in the School of Computing and Augmented Intelligence, part of the Ira A. Fulton Schools of Engineering at ASU, has received funding through the Geospatial Innovation for Food Security Challenge, or GIFS Challenge, an initiative from Taylor Geospatial that supports geospatial artificial intelligence, or GeoAI, solutions for global food security.
Tárano and her collaborators in the Kerner Lab will develop a system called “Promptable, Uncertainty-Aware GeoAI for In-Season Forecasting of Food Assistance Infrastructure in Conflict Zones.” The project aims to help humanitarian organizations identify early signals of instability in crop production and food assistance infrastructure in regions affected by conflict.
“Our goal is to help our partner organizations prevent famine and reduce acute food insecurity by identifying early signals of instability in food systems, enabling them to deliver humanitarian aid to the right places at the right time,” Tárano says.
The project brings together researchers and practitioners from ASU, the University of Maryland, Washington University in St. Louis, NASA Harvest and the Famine Early Warning Systems Network, or FEWS NET. Together, the team will develop and evaluate open-source tools that can help decision-makers better understand rapidly changing conditions in some of the world’s most vulnerable regions.
Unlike traditional geospatial analysis tools, which often require specialized technical expertise, the new system is designed to make satellite-based insights more accessible. Users will be able to ask questions in plain language, such as whether agricultural fields appear prepared for planting or whether signs of disruption are emerging in key food-producing areas. The system will then generate data-driven assessments while clearly communicating the uncertainty associated with those predictions.
That emphasis on transparency is central to Tárano’s research.
Her work focuses on developing human-centered artificial intelligence systems that help people make informed decisions in complex environments. By combining advances in machine learning with user-centered design, she seeks to ensure that AI tools are understandable, trustworthy and useful for the people who rely on them.
The project also reflects the broader mission of the School of Computing and Augmented Intelligence, which advances computing technologies that address real-world challenges across areas such as sustainability, health, education and public service.
“Food security is a critical global issue,” Tárano says. “We want to create AI systems that not only generate insights but also help decision-makers understand what those insights mean and how confidently they can act on them.”
The project draws on the expertise of Assistant Professor Hannah Kerner, whose research focuses on applying machine learning and geospatial AI to challenges in agriculture, climate resilience and disaster response. Her work on satellite imagery analysis helps provide the technical foundation for tools that can rapidly assess conditions in conflict-affected regions.
“This project will bring academic research on geospatial foundation models and embeddings into the real world, making them usable in rapid response scenarios for assessing conflict impacts on agriculture,” Kerner says.
Implementation partners NASA Harvest and FEWS NET bring extensive experience translating Earth observation data into actionable information for governments, humanitarian organizations and agricultural stakeholders worldwide.
“Decision-makers often have to assess food security risks with limited and delayed information about what is happening on the ground,” says Inbal Becker-Reshef, director of NASA Harvest. “By generating more timely and transparent information, it will help address critical gaps and support organizations working to anticipate emerging food security risks.”
Ana M. Tárano (left) speaks with farmer Negus Manna during a field visit on the island of Lāna’i, Hawai’i. Tárano’s research focuses on developing human-centered AI systems that translate satellite data into practical information for agricultural and food-security decision-makers. Photo courtesy of the Kerner Lab
The research team plans to test the technology in conflict-affected regions including Sudan, Ukraine, Syria and Haiti. While those locations provide immediate opportunities for evaluation, the researchers say the framework could ultimately support food security monitoring efforts in vulnerable regions around the world.
The project aligns closely with the goals of the GIFS Challenge, which seeks to accelerate the development of practical GeoAI tools that can move beyond research settings and into operational use. By bringing together researchers, humanitarian organizations and implementation partners, the initiative aims to shorten the path between scientific discovery and real-world impact.
For Tárano, the project’s success will ultimately be measured by its ability to support organizations working to prevent hunger and improve lives.
“When humanitarian organizations have better information, they can make better decisions,” she says. “Our goal is to help ensure that critical resources reach communities before food crises escalate.”