News

One of the biggest misconceptions about AI in public health is that it is a magic bullet. Something that can instantly solve deeply entrenched systemic issues like health inequity or data gaps. AI is a powerful tool, but only as good as the data it’s trained on and the context in which it is deployed. People assume AI will generate perfectly objective decisions, but if the data is biased, incomplete, or unrepresentative, it can perpetuate disparities rather than fix them. Some think AI can replace public health employees, but it is meant to augment and not replace their judgment.

Another thing is that AI can hallucinate and include fabricated references, invented facts, and potentially dangerous instructions in its response to a prompt. Always fact check AI output before using it!

• AI is a helper and not a replacement of employees. Public health professionals bring experience and specific knowledge of the community they serve, which are things AI cannot replicate.
• AI can augment decision-making and should not be a substitute for it. Think of it as a super-fast assistant that does not sleep but still needs supervision and oversight.
• AI can invade privacy or be misused. These are concerns that require AI governance and a usage policy throughout a health department.
• When the public health workforce understands how to strategically engage with AI, it becomes a transformative ally rather than just another tech trend.

• First, understand your current AI Position and how much AI is currently being used. PHF developed the Stages of Agency AI Adoption as a tool we use with our clients to help them get the big picture. We also use our POAR matrix to help our clients understand their current AI Position, Opportunities for AI, how to Adopt AI without disruptions, and what Results they want to achieve with AI.
• Make the AI policy simple and straightforward by drafting easy-to-understand guiding principles. You do not have time to develop a 73-page policy covering everything. Cover what matters, implement it, and revise as you go along. These principles should guide AI development, deployment, and governance.
• Define why you are integrating AI into the health department. Is it for disease surveillance? Resource allocation? Community health forecasting? To support team-based problem solving? Other?
• Assess Data Readiness. Review current data systems and determine if they are secure, clean, and representative of the community you serve.
• Technology changes fast, so treat your AI policy as a living document and revisit and revise it on a regular basis.

If public health organizations avoid AI altogether or delay implementing AI policies, they risk missing out on transformative opportunities to help their organization and the communities they serve. Such as:

• Missed chances for early detection and response. AI can flag emerging disease clusters, forecast outbreaks, and identify anomalies in real time. Without it, response times are slower, costing lives and straining resources, especially in fast-moving crises like pandemics or environmental disasters.
• AI can help optimize staffing, supplies, and interventions by analyzing patterns that humans may overlook.
• Public health organizations collect enormous amounts of data, but AI is what makes it actionable quickly. Not using AI means drowning in spreadsheets, while others using AI are turning those insights into action.
• It is not about jumping on a bandwagon but thoughtfully engaging safely and securely with a powerful tool that can enhance every facet of public health. If AI feels intimidating, starting with simple policies and small pilots is the safest way to learn, adapt, and lead.

• Use a licensed chatbot such as Microsoft Copilot to see how it can enhance communications, give you new ideas on a problem, write memos, and more.
• Be careful using open-source AI since in the past this was a hacker’s highway, and you do not know what you have downloaded or how it can impact your IT systems.
• AI-powered productivity suites using tools such as Microsoft Copilot in Word, Excel, and Outlook can assist with writing and summarizing.
• There are many graphic AI packages that can help enhance presentations.
• Start small to understand the power of AI and then expand its usage.

Expert guidance can help you to:

• Develop secure chatbots to protect organizational data.
• Provide data protection by designing AI systems that protect sensitive information.
• Understand the cost of implementing and sustaining AI in your organization.
• Recognize the pitfalls of using open AI models or buying off-the-shelf AI models.
• Develop an AI Adoption Team to determine what is best for the organization and assess ongoing costs.
• Integrate AI quality improvement teams and performance management activities.
• Train staff in the use of AI. Organizations need to decide who can use AI, where they can use it, what type of AI they can use and not use, and then show them how to use it.
• Address privacy and legal concerns. AI can cause problems with privacy laws, so organizations need legal and policy experts to ensure the organization stays on the right side of regulations like HIPAA or local governance rules.