What AI Can (and Can’t) Do for Public Health
A Post-Event Recap from Metopio and NACCHO’s Webinar
AI is everywhere, and it’s only getting louder. But beneath the buzzwords and billion-dollar headlines, public health leaders are asking the big questions: What’s real? What’s safe? And how do we make sure our use of AI is helping the communities we serve?
That’s exactly what we explored in our latest NACCHO-hosted webinar, What AI Can and Can’t Do for Public Health, featuring Metopio CEO Will Snyder and CTO Jonathan Giuffrida. In case you missed it (or just want a refresher), here’s a quick recap.
AI Is Booming — But Public Health Must Set the Terms
“We don’t have the luxury to chase the hype or to ignore it,” said Will Snyder as he kicked things off. “We have to ask the hard questions in front of public health right now, like how do we make sure AI strengthens our work rather than distracts from it — or causes misinformation or distrust?”
With billions in global investment pouring into AI and only a fraction of public health agencies ready to deploy it, the pressure is on to be discerning. And as we all know, public health isn’t just another sector. This work lives at the intersection of data, policy, and people’s lives — it’s about trust, context, and equity, so having the right tools matters!
What AI Can Do: Save Time, Surface Patterns, and Support Strategy
So, when choosing an AI tool, what should you look for? What’s AI actually good at? Will and Jonathan highlighted a few sweet spots:
Speeding up repetitive work: Analyzing transcripts, building dashboards, and generating captions are all great examples.
Drafting CHA/CHIP reports: Let AI take a swing at a first draft based on user data, past assessments, and national datasets — while keeping humans in the driver’s seat before next steps or publication.
Simplifying community engagement: Consider offering an AI-enabled chatbot that can interpret community questions, pull real data, and cite sources.
Freeing up capacity: Today, public health professionals need to focus more on primary data collection and strategy, not formatting charts.
As Jonathan put it, “AI can do the heavy lifting — like analyzing 100 hours of interviews in seconds — but it still needs your judgment to prioritize what matters most.” And that brings us to our list of things AI cannot do.
What AI Can’t Do: Replace Human Expertise or Local Nuance
Despite the power of AI, there are real limits. This webinar outlined four:
It can’t replace ethics or experience
It’s not perfectly accurate
It lacks place-based knowledge
It can’t set public health priorities
“A generative model has no idea what it’s like to live in your county,” Jonathan reminded the audience. “And if you let it operate without constraints, it’ll happily invent fake community members or offer a report of an entirely incorrect city just to make you happy.”
Guardrails, Frameworks, and the Role of Policy
The webinar also covered policy best practices, including:
Keeping AI away from raw data (especially PHI)
Training models only on vetted public sources and human-reviewed text
Ensuring transparency, human oversight, and feedback loops at every step
If your team doesn’t yet have an AI policy, “you’re not behind,” said Will. “But now is the time to start building one that clarifies where and how AI can be used.”
Join us for a practical, no-fluff conversation about building AI policies that are credible, flexible, and actually usable in public health work. AAron Davis (Director of the Center for Public Health Initiatives at Wichita State University's Community Engagement Institute) and Tatiana Lin (Director of Business Strategy & Innovation at the Kansas Health Institute [KHI]) will break down KHI’s AI Policy Template, an emerging standard that’s already helping local and state agencies navigate issues like privacy, bias, human oversight, and community trust.
Bonus: NACCHO members have access to a free Health Atlas from Metopio — a great entry point for exploring AI-powered community dashboards. Check out how the City of Chicago has set up their Atlas!
What’s Next?
As AI grows and changes, Metopio is focused on three near-term opportunities:
Automating CHA workflows
Launching conversational analytics for epidemiologists
Making visual storytelling easier than ever
The goal, as always, is to keep humans in charge while making sure they’re not buried in spreadsheets.
“AI is not here to replace public health,” said Jonathan. “It’s here to help you do what you already do — faster, clearer, and with more time to focus on the work only you can do.” Missed the live event?