1156: PHIG Impact Report: AI in Public Health

JOHN SHEEHAN: 

This is Public Health Review Morning Edition for Tuesday, June 30, 2026. I'm John Sheehan, with news from the Association of State and Territorial Health Officials.

 

On today's PHIG Impact Report, artificial intelligence is already finding its way into public health agencies, from drafting communications and translating materials to analyzing data and identifying emerging health threats. The challenge for leaders is no longer whether to use AI, but how to implement it responsibly. Today, we'll hear from Tatiana Lin and Shelby Rowell of the Kansas Health Institute. Tatiana is director of business strategy and innovation. Shelby is director of the Region Seven Public Health Innovation Hub. They'll discuss how health departments are using AI to improve efficiency, support community engagement, and expand workforce capacity, while addressing concerns about privacy, bias, transparency, and governance.

 

Tatiana, let's talk about the Kansas Health Institute's work with AI, and specifically how PHIG is helping support it.

 

TATIANA LIN: 

So, in 2023, we were very excited to launch some of the work in artificial intelligence. It all began with the whole interest and excitement in AI, and we wanted to make sure that public health organizations, and specifically health departments, feel supported and not left behind in this conversation. We also recognize that AI holds potential to inform and influence the public health space, and it's really important that public health professionals had access to guides, practical resources to ensure they can meaningfully participate in the discussion. So, through this partnership, we, at the Kansas Health Institute, a part of the Region Seven Public Health Infrastructure Hub, we had the opportunity to develop, back in 2025, an AI policy template and guidance. And we'd done this through the Public Health Infrastructure Grant, but also in collaboration with Health Resources in Action, and with the State University Community Engagement Institute. So, the purpose of that resource was really to establish AI governance framework for organizations, so they can easily, more easily develop their AI policy. And we started there because a lot of organizations had a question about: How do you develop your AI policy? It was really difficult to do because they didn't know what to include in their policy, how to structure the policy, and also what are the most important provisions. We also know that public health organizations have a lot of responsibilities for the communities to protect sensitive data and to maintain public trust. So, really starting with the AI governance structure was really important, and PHIG played a really important role in this, because it helped us to, first of all, to even do this resource or start this work. Then it really helped us to provide trainings and capacity building opportunities to the communities and jurisdictions in that space, and also helped to inform resources as well.

 

SHEEHAN: 

And you mentioned that you were hearing a lot of interest in AI and getting a lot of questions. Why is it important for public health agencies and organizations to really jump on that bandwagon and understand that AI is kind of here, and you know, implementation is going to happen no matter what?

 

LIN: 

Great question. So, as you mentioned, public health organizations already encounter AI in a lot of spaces; it could build search agents or just AI-embedded features in already existing tools, even Zoom has AI assistants. So, the question is no longer if AI is present in public health, because it already is. It's like, how can we really leverage it as value-added and keep the kind of responsible AI applications in mind? So, there is that whole: "AI holds potential for public health," and it's not there to replace human expertise, but it can really maximize what we can potentially do, because what AI can really do is to process enormous amounts of information very quickly that people, it's harder to achieve alone without technology. We know that public health, specifically AI, can process large amounts of data faster than humans can. It can help to identify trends, it can help to identify emerging issues. We're thinking about this literature, environmental scans, where we typically spend so much time looking through articles. Now, we have this opportunity to not just select a few articles to review; we can do this, you know, to all the literature that exists up there. We can also monitor multiple informational sources at the same time. It has the power to help us translate the documents in different languages, and it's so important for public health, because we serve diverse audiences. So, there's all of those different potential uses if we apply or use AI responsibly, and that's where, also a very important point here, is what does it mean to use AI as a value-added tool? Because we know that AI also has some ethical considerations, so, how do you balance that? How do you balance what we know about environmental impacts of AI versus value-added of artificial intelligence?

 

SHEEHAN: 

Yeah, and you know, using AI responsibly is of course of paramount importance, but how can that responsible use lead to better serving communities?

 

LIN: 

I think that's one of the primary goals, because sometimes when we talk about AI, we talk a lot about efficiency, because that's where, kind of, it's clear what that's where the benefit could be. But it's really important to say, to go kind of beyond efficiency and think more about quality, about the other values that AI can bring, and specifically thinking about the community and how health departments can maximize [that] their value is responsible. I'm thinking about this as it could serve as a force multiplier for public health. So, it can do— a lot of health departments have limited resources, and AI can really help with that to supplement those resources and help staff to focus on, I would say, human-centered work, where we can really work with communities to build trust, you know, engage residents and address local needs. So, when implementing AI responsibly, I think it can help health departments to improve access to services, as well as to identify the needs of the communities faster. For example, if we have residents that need information and resources [in] multiple languages, AI could help to provide that translation. It'll be really good to verify it with bilingual and trusted community liaisons as well. It can help to analyze community feedback very quickly so we don't spend as much time on analysis, but can quickly, more quick[ly] jump into action into actual implementation stages as well. So, really that reduction in the time spent on information processing tasks, and really spending more time on partnership development in, kind of, community engagement, where human judgment and relationship-building are very critical.

 

SHEEHAN: 

So, Shelby, what questions do you get from public health organizations when it comes to implementing for use cases?

 

SHELBY ROWELL: 

I think what we're hearing most consistently is a desire for practical time-saving tools that make their day-to-day operations smoother. Public health departments are stretched thin right now, and often the use cases that we're seeing are mostly administrative or operational. So, things like drafting standard operating procedures, turning a short event description into a press release or a set of social posts, and summarizing public health documents and streamlining internal communications. I think, interestingly, what we're also seeing right now is kind of the desire to look beyond some of these administrative tasks, so seeing things that are a little bit more innovative or cutting-edge. So, interest in things like using AI to help catch improper or false Medicaid claims denials, or using AI as a tool within surveillance. We see a use case right now at the University of Southern Florida, where they've taken AI image recognition with citizen-submitted photos, and that's helped them identify disease-carrying mosquitoes, so they were able to identify a disease-carrying mosquito in Madagascar, which kind of led to their surveillance and activities after that. So, I think departments are also wanting to see what are some of those use cases that they may not be able to implement right now, but they want to in the future.

 

SHEEHAN: 

Wow, sure. Okay, let's go to the other side, then, and start pumping the brakes, and say, what are the questions and concerns around policy and around governance of AI?

 

ROWELL: 

I think right now, what we're seeing is that they don't want to just start from a blank page. AI is an area of technology that's constantly evolving, and it can kind of be a little bit unsettling or scary to dive into the policy side of things if you don't understand what AI is and what AI can do. So, having, you know, a resource that they can take and adapt rather than build from scratch, I think that's something that we're seeing a lot of health departments say they need. And even outside of health departments, that's what we're seeing as a need within kind of the broader health field as well.

 

SHEEHAN: 

And you've worked with public health departments, such as the one in Colorado. What needs did you identify that they were trying to shore up, and what did those collaborations look like?

 

ROWELL: 

Yeah, so I think broadly, even outside of Colorado, they're kind of consistently wanting plain language guidance on the things that, I don't know, maybe keep us up at night in terms of being within public health and using AI. So, things like data privacy, bias mitigation, transparency, and who's accountable when something goes wrong. I think with Colorado, we're actually more so engaged with the Colorado Association of Local Public Health Officials. So, they came to us with a technical assistance request, where they really want something for their members that's adaptable, practical, so, an AI policy use template that can be applicable to a variety of different types of health departments, from your largest health departments in urban areas all the way down to rural and frontier health departments that may have one or two staff members. Those policies are going to look very different, because their operational realities are different. So, with that collaboration, we've taken a step back to see, what are some of the needs that would be applicable to all of the health departments? And then, also building in the template, different areas that can be adapted based upon those local realities.

 

SHEEHAN: 

And so, Tatiana, you helped some agencies in Colorado and Guam navigate their own implementations of AI tools. I'm wondering, what kind of needs that you saw them addressing, and how do those implementations shake out?

 

LIN: 

Yeah, so I love working with different jurisdictions. It was such a wonderful opportunity through the Public Health Infrastructure Grant to engage with the Guam Department of Public Health and Social Services. Their specific ask was to help them think through how to develop their AI governance in the AI policy. They recognize the potential for AI, but they also wanted to have that, kind of, balanced approach to make sure that they're aware of what they should safeguard and also how they can really start that implementation. So, we worked with them to develop their AI policy. What we really appreciate [about] the process [is] that they have a really representative team that came from their department to be really, co-lead those efforts together, where we provided facilitation and support, and they provided kind of a thought process in the context of their department. Also, what we helped to do through the Public [Health] Infrastructure Grant is to provide them with a survey tool that they were able to administer: the department collect[ed] information from their staff, what their staff thought about AI, what self-rated capacities they have, what concerns they have, and through that survey was instrumental in actually informing the policy. So, we helped to bring that information into the policy discussions and ultimately that informed quite a few sections of the policy that they have developed. So, the biggest win there was, right now, is [to] figure out how you create that participatory approach to development of AI framework, AI governance framework, and infrastructure, and engage staff and create that as a foundation for future AI implementation.

 

SHEEHAN: 

Using PHIG funding: you've developed AI implementation workshops and trainings. What [do] those trainings and workshops focus on?

 

LIN: 

It's some other AI work we do at KHI that also is our favorite component, because it allows us to really understand where the needs of the public health community [are] and how we can really tailor [the] trainings and workshops to their needs, but our workshops are multidimensional. A lot of them focus on AI literacy and capacity building. Then we have workshops about practical implementation, where we typically discuss where some use cases are in public health [and] how you really leverage generative AI systems, how you do this effectively. The term that's really popular right now is 'prompt engineering,' so how do you really do the prompt engineering with a generative AI systems? Also, we focus a lot on ethics, so we really appreciate the opportunity to walk through what are some potential considerations, and how you really mitigate that through individual actions, organizational actions, and also policy. And lastly, some of the workshops focus on, again, this AI governance structure building, and how do you build your AI policy as well. And so, sometimes with organizations we offer through this continuum, they start with capacity building, and then we kind of end up with a policy, but [it] sometimes depends where they are in their AI journey; we do what meets their needs,

 

SHEEHAN: 

And so, Shelby, when you're sort of preparing the menu of options for a public health department or a public health leader, how do you decide what to present to them as these are [some] things you could do versus tailoring it to all [of] the things they could do?

 

ROWELL: 

Yeah, so the work within this new kind of reactive TA project, it kind of builds upon two things, as we were building the AI policy templates for CALPHO. The first is our resource that we developed with [the] Wichita State University Community Engagement Center, as well as Health Resources in Action, and that's the "Developing AI Policy for Public Health Organizations," a template and guidance. So, that's kind of the first base of where we were able to pull out information from that document and put it into a more ready-to-use template. The other part of it is our experience working with the Guam Department of Public Health and Social Services on their own AI policy. In the conversations that we had with them about what would they want to include within their AI policy, we also incorporated that within the CALPHO template. So, there's certain things within the new template document that we do recommend as as core principles, so things like making sure that we're being transparent in our AI use, ensuring that we're trying to mitigate bias as much as possible, human oversight; those are the things that are really core to any AI policy. But there's also specific areas that go into, you know, how does this go into operations? Things like, how do you identify what are the risk categories for different AI use cases, and what are the appropriate types of artificial intelligence that you can use as you're, you know, incorporating those different low-risk, medium-risk, and high-risk use cases.

 

SHEEHAN: 

And what role do you see public health leaders playing in shaping these conversations around how their departments should be using AI?

 

ROWELL: 

I think right now we're moving into a phase where leaders are really setting the policy and leading, kind of, the policy development creation for that, but the workforce is really the folks that are on the ground that are using AI for the different areas. So, I think leadership's role is to keep building the governance and training and accountability structures, and that leads to the experimentation where AI use is happening in a safe environment or a safer environment, rather than banning AI outright or letting it run without any oversight. So, I think that's the role that leaders have to keep playing. So, it doesn't just go away once a policy is signed. It's really an iterative process.

 

SHEEHAN: 

Have there been any surprises in real-world use cases, or have there been any lessons or takeaways from how people are actually using this technology?

 

ROWELL: 

What surprised me is just how much of an appetite that there is within health departments. I think we're a really unique and special community of leaders, and the fact that when we're implementing a new technology or a new program, we want to make sure that you know it aligns with our core values. So, I think that's been, in a way, not surprising, but that's what excites me; is when we're talking about public health and their implementation of AI, they want to do it in a safe way. They want to make sure that community members are being protected and their data is being protected. So, we've seen that with our work doing trainings all across the country with our partners at [the] Wichita State University Community Engagement Institute; we've done over 90 trainings in the past three years, reaching just over, I think, 9,000 individuals. So, that's not just a one-time spike of curiosity; it's really a sustained engagement and wanting to deepen their understanding of AI and learn as much as they can to make sure that they're doing it in a way that aligns with their core values.

 

SHEEHAN: 

So, Tatiana, outside of ethics or guardrails, what kind of challenges do you run into when helping agencies implement their own, their own versions?

 

LIN: 

It's interesting, I think, when we started doing AI capacity building workshops, because AI was, although it's not new, you know, AI goes back to, you know, [the] 1950s, it was more newer for public health community, especially in that new version, where it's more powerful, where it's more accessible; we have so many different tools. So, initially those workshops were really just, "Tell us about AI," you know, "What do we need to know? What's happening?" And as we're more in that AI progression, I think the technical assistance requests get more nuanced there. And it's great to see that public health organization want to really tailor it to their needs. So, questions come about [like] what the AI ready infrastructure looks like, what do we need to know about this, or if we have the certain capacity, how do we really pick the right vendors or the right AI tools. And so, I think that some of the challenges are that there is no one size fits all, there's such a different level of readiness, such different rail of capacity across their organizations. So, being really able to tailor that to their needs. I think another one is just really understanding and taking some time to understand what the lay of the land is, because the AI tools are moving so fast, the infrastructure is moving so fast. So, just keep[ing] up with the research and keep[ing] up with best practices is something else that's really important for TA providers as well.

 

SHEEHAN: 

And as you say, the technology is moving and evolving and growing so quickly. Do you see that as creating more challenges for smaller agencies or more opportunities?

 

LIN: 

That's a fascinating question. I think it's definitely both. On one hand, we do want to have a lot of different tools and systems that can meet different needs of organizations, but I think what would be really important to understand, which ones of them are really tailored to the public health needs and which ones of them have the public health capacities, you know, as well, in mind. And so, that really creates an opportunity at the same time, because it's moving so fast, and there are, you know, literally 1,000s of different tools and systems, it's really hard to understand how to pick them, how to engage in conversations with the vendor, how identify and assess the technical characteristics of the systems, their security profile, and who is within the organization are positioned well to do that. So, because for a lot of, I believe, IT departments, they are also building capacity in understanding that AI landscape, in being able to assess that as well.

 

SHEEHAN: 

Tatiana Lin and Shelby Rowell of the Kansas Health Institute. Tatiana is director of business strategy and innovation. Shelby is director of the Region Seven Public Health Innovation Hub.

 

This has been the PHIG Impact Report, part of Public Health Review Morning Edition. I'm John Sheehan for the Association of State and Territorial Health Officials.