University of Hawaiʻi Gets $2M to Test AI Early Warnings for Food Risks

The University of Hawaiʻi at Mānoa has received a $2 million National Science Foundation award for AI-enabled tools intended to identify environmental threats to food-production systems before they become larger problems. The work is part of a four-year, $4 million collaboration with the University of Nebraska–Lincoln.

From outbreak response to environmental warning

The proposed system addresses a timing problem. Principal investigator Tao Yan cited Cyclospora-linked illness investigations, where the problem is recognized after people have consumed food. The new tools are meant to look for environmental threats before they grow into broader food-safety problems.

That is an ambition, not a demonstrated result. Yan has described the goal as early-warning technology that identifies emerging risks before they reach critical levels. The project is intended to improve food safety, protect animal health and strengthen food-production systems.

AI interprets the samples; it does not replace them

Researchers plan to combine AI with environmental sampling, metagenomics and high-resolution mass spectrometry. Metagenomics examines genetic material from environmental samples, while high-resolution mass spectrometry identifies chemical compounds. The models are meant to interpret those biological and chemical inputs for potential risks in aquaculture, livestock and agricultural systems.

The distinction is important: the research design relies on physical samples and laboratory analysis, with AI used to interpret complex data. Whether that combination can produce useful warnings under real production conditions is what the planned initial trials will test.

The first test sites narrow the scope

Initial testing will focus on aquaculture and beef cattle production systems, rather than every type of food production. The four-year project began September 1 and is led by Yan, director of the University of Hawaiʻi Water Resources Research Center and a professor of civil, environmental and construction engineering.

Work beyond the trials

• The collaboration plans outreach with industry partners, regulators and communities to encourage adoption of the technologies.
• It also plans research opportunities for junior faculty members, graduate and undergraduate students, and K-12 participants.

The project’s evidence will come from those trials, not the grant itself. Its core question is whether AI can turn genetic and chemical signals from production environments into warnings early enough to help protect food, animals and the people who consume the food.