An auto body shop repeatedly violates air-quality rules next to a row of homes. A bus route stops short of the nearest grocery store. A longtime public health worker knows from years in the field which neighborhood patterns never make it into a spreadsheet.
For Katie Stebbins, executive director of the Food & Nutrition Innovation Institute at the the Gerald J. and Dorothy R. Friedman School of Nutrition Science and Policy at Tufts University, such scattered pieces of information can be crucial to understanding public health in a community.
But too often, Stebbins says, these facts remain just that: scattered.
Artificial intelligence, she believes, has the potential to help public health professionals discover and use fragmented knowledge to make better, more transparent decisions. “There’s a whole lot of nuanced, local, almost parcel-level data that we’re just not yet capturing,” Stebbins says. “AI can help, if we use it in the right way.”
On June 12, Stebbins will share her thinking in “AI in Public Health: The Value of Learning a New Language,” a panel at the Connecting Sectors, Connecting People: Tufts Public Health for Critical Impact symposium. This event reflects a distinctly Tufts model of public health, says Thanos Zavras, one of the symposium's organizers. Zavras is the chair of the Department of Public Health and Community Service and associate dean for community and global affairs at Tufts University School of Dental Medicine (TUSDM).
“Tufts public health is purposefully outward-facing and constantly cultivating intersectoral partners to expand opportunities for impactful solutions,” says Zavras, who holds the Delta Dental of Massachusetts Professorship in Public Health and Community Service at TUSDM. The symposium, he adds, is meant to show how connecting people from different sectors can energize diverse skillsets to focus on population health.
That broad, cross-sector approach is central to how Stebbins thinks about AI.
Before joining Tufts, she worked in economic development in Western Massachusetts, including in Springfield and Holyoke, where she saw how public-health challenges could limit workforce opportunity and weaken a community’s economic potential. The experience taught her to look for links among health, work, environment, infrastructure, and community life—the kinds of connections that AI may help public health professionals see more clearly.
Here, Stebbins discusses what it will take to make AI useful, trustworthy, and accessible in public health.
Public health is facing growing skepticism and political pressure. Where do you see the greatest opportunity for AI to help the field rebuild trust and relevance?
Too often, we treat public health as separate from economic development and workforce development. But in the communities where I worked, those issues were inseparable. If people are not well enough to work—especially in low-resource communities—that affects families, employers, and the economic potential of an entire place.
AI can help by making patterns visible across systems that are usually managed separately: health, housing, transportation, food access, environmental risk, and workforce opportunity. To rebuild trust and relevance, public health has to become more useful and more visible in people’s daily lives. AI should help practitioners ask better questions and understand local conditions more fully—and ultimately make better decisions that people can see and benefit from.
The title of your panel frames AI as a “new language.” What does that metaphor mean to you in practice?
Early in my career, I redeveloped brownfields—polluted, abandoned industrial properties—and worked with residents to understand technical reports on geology, toxicology, chemical discharge, and other risks. The goal was not to explain the experts’ language but to translate it into terms the neighborhood could use in its own planning.
That practice of translation is central to community work. Economic development has a language. Land-use planning has a language. Public health, medicine, and environmental science all have languages. The same is true when it comes to AI.
If community leaders are going to use AI well, they need more than access to the tools. They need to know how to ask the right questions and develop enough fluency with AI to understand what it can clarify, what it may flatten, and where human judgment still has to lead.
AI is often discussed in highly technical terms. What do you think public health leaders most need to understand—not about the technology itself, but about how it changes the way problems are framed and solved?
Public health practitioners have a lot of data, but some of it is organized and some of it is not. Some may be in the cloud, and some may still be on servers in messy closets. And then there is the knowledge held by people who have worked in a place for 40 years—the kind you get only by riding along through a neighborhood with someone who can tell you what has changed and what patterns hold significance.
AI changes the framing because it asks us to think about what information we have, what we can know, what we cannot know, and what questions we should ask. It also requires public-health leaders to understand what they are being sold. Government agencies often have limited resources, and contractors may promise that a new tool will deliver everything they need.
Practitioners need to know how to evaluate those promises so AI does not just look like magic.
At the Food and Nutrition Innovation Institute, you focus on moving ideas from research into real-world solutions. How is AI changing the pathway from academic discovery to public-health impact?
The food system is infinitely complex. Public health, nutrition, agriculture, food as medicine, supply chains, infrastructure, genomics: there are so many people in so many places doing so many things, and everyone is raising their hand and saying, “My approach is right.”
At the institute, AI has been a way to organize that complexity. It allows us to step back and ask: How does this piece of work relate to what’s happening in nutrition genomics? How does it relate to supply-chain delivery infrastructure? How do we connect these dots so we can have a more intelligent conversation? Answering questions like those might help us better understand how academic discovery can translate into impact.
Looking ahead five to ten years, how do you hope Tufts—and public health more broadly—will be using AI differently to respond to global challenges in health, nutrition, and resilience?
Public health will have to be recentered as a societal pillar. AI may help by showing how systems that now operate separately could work together to create safer environments faster.
“Faster” is a key word. Speed will become increasingly important as communities face more complex disasters and public health challenges. We will need systems that can analyze large amounts of information, make the data visible, and support more transparent decision-making quickly. I hope AI will help there.
As far as Tufts in particular goes, I hope that we push ourselves to look beyond one- to five-year cycles and ask what foundation we need to build now not for 2027 but for 2046 or 2050.
|
|