Why and How Kansas Public Health Could Be Key in Shaping a Statewide AI Roadmap

• Home
• Articles

Why and How Kansas Public Health Could Be Key in Shaping a Statewide AI Roadmap

Transforming Public Health for the 21st Century: Bridging Theory to Practice is a blog series that will explore the challenges and opportunities faced by the public health sector and will introduce innovative ideas and unique voices to support transformation. Sign up here to receive these summaries and more, and also follow KHI on Facebook, Twitter, LinkedIn, and Instagram. Learn more about the series on our archive page. Please feel free to share your feedback or suggestions with us by emailing [email protected].

Artificial Intelligence (AI) has existed for many years, but it rapidly became a familiar term when OpenAI launched ChatGPT in 2022. This development sparked widespread interest in exploring the capabilities of generative AI (GenAI) and other forms of AI technologies. However, many organizations approach this exploration with caution or refrain from it altogether, often due to the absence of robust guidelines and policies. The development of guidelines has been gradual because organizations have been hesitant about how to proceed without readily available technical expertise and a comprehensive understanding of AI's potential ethical, social and legal implications - both known and unknown. Other contributing factors include the evolving landscape of AI-related research, as well as federal and state regulations. Additionally, developing comprehensive AI guidance requires resources that may compete with other priorities.

Waiting for these issues to be fully resolved before acting may not be the optimal strategy, as inaction could lead to falling behind in leveraging AI's potential, becoming reactive rather than proactive in managing potential risks and missing critical opportunities to integrate equity and ethical considerations into the AI landscape. As the field of artificial intelligence evolves, steps could be taken at various levels - from creating organizational guidance for AI use and safe experimentation to developing action plans or roadmaps at the state level. Effective guidance likely will require striking a balance, as overly high-level or restrictive approaches might hinder exploration and limit positive outcomes, while overly permissive guidance may fail to address ethical, social and legal considerations.

This blog focuses on the reasons why public health should be at the forefront of AI dialogue at all level, highlights several AI-related efforts in Kansas and provides ideas for next steps.

AI Roadmaps and Legislation Across the U.S

Although guidance and research are still emerging, a growing body of resources from across the country can be used to inform the development of localized strategies and frameworks. For instance, several roadmaps and action plans have been developed across the country. The 2023 New York City AI Action Plan and the 2024 Los Angeles A.I. Roadmap are prime examples. In addition, at least 13 cities across the country, including Austin, Texas; Boston, Massachusetts;Lebanon, New Hampshire and Tempe, Arizona, have also created their own AI guidelines or policies. At the state and territorial level, according to the National Conference of State Legislatures (NCSL), during the 2024 legislative session, at least 40 states, Puerto Rico, the Virgin Islands and Washington, D.C., introduced AI-related bills. These bills focused on areas such as addressing potential algorithmic bias, consumer protections, educational integration, government oversight, task force establishment, legal clarifications, policy frameworks, public safety, crime and AI in business and workforce development. Additionally, six states, Puerto Rico and the Virgin Islands adopted resolutions or enacted legislation. In 2023, Kansas was one of several states to adopt a generative AI policy for executive branch agencies. The Association of State and Territorial Health Officials (ASTHO) developed an interactive map that captures the development of state and territorial legislation related to public health systems. Other organizations, such as MetroLab, a nonprofit, and the National Association of Counties (NACO), have launched initiatives that focus on GenAI for local governments.

AI and Public Health

AI has been applied across various industries, including public health. The Centers for Disease Control and Prevention (CDC), for example, uses MedCoder, a system that utilizes natural language processing and machine learning to code causes of death, automating nearly 90 percent of records compared to the previous rate of less than 75 percent. Similarly, the Chicago Department of Public Health has used AI to identify children at high risk of lead poisoning, prioritizing home inspections through historical data analysis. To support real-time surveillance, AI can be used to analyze social media to monitor public sentiment, detect emerging health trends and identify potential disease outbreaks. By tracking keywords and discussions, AI could enable more rapid and effective responses to emerging health issues. However, to optimize the benefits of AI, it's essential to ensure models are unbiased and address data privacy concerns. Understanding potential sources of bias in AI algorithms and developing strategies to address them can help reduce the risk of reinforcing existing inequities.

Beyond this, AI holds potential for a range of applications. AI can be leveraged to assist in generating code or checking it for errors. It also can support community health needs assessments by performing tasks ranging from creating survey questions to summarizing results. Generative AI tools like ChatGPT can further enhance the public health workforce by aiding in administrative processes, creating health communications, drafting press releases and generating educational materials and training resources. In the use of AI tools, it is essential to maintain a human-centric approach to ensure ethical oversight, accuracy, and the responsible application of AI in public health. Additionally, it is crucial to recognize the environmental impact of AI technologies, as well as address copyright concerns related to AI-generated content.

Public Health's Role in AI Roadmap Development

Public health is well-positioned to be at the forefront of AI roadmap development due to its strong commitment to ethics and equity, its focus on enhancing population well-being and its efforts to reduce health inequities and disparities. With its expertise in data-driven decision-making, public health can support the development of equitable and transparent AI technologies to avoid exacerbating health inequalities. Additionally, its interdisciplinary approach and proactive risk management make it uniquely suited to identify ways to integrate AI across initiatives while addressing potential concerns related to AI systems, such as data privacy and algorithmic bias. Furthermore, the sector can play a critical role in raising important issues in the AI space, such as the impact of the digital divide and rural considerations, on the equitable implementation of AI.

Even if public health may not have all the expertise needed in AI, it could serve as a convener for discussions, engage with the private sector to inform the development of AI systems and assist in creating equitable training programs and AI guidance.

Kansas Initiatives

Community of Practice

The Center for Public Health Initiatives (CPHI) and the Kansas Health Institute (KHI) have partnered with local health departments (LHDs) to advance the integration of AI in public health across Kansas. In September 2023, CPHI and KHI launched a 12-month Community of Practice, funded by the Kansas Department of Health and Environment (KDHE), to support this effort. The collaboration aims to leverage AI to enhance operational efficiencies and address the diverse needs of LHDs. It has created a collaborative space for exploring the use of generative AI in daily operations and creative problem-solving. While the full range of AI applications is still being explored, encouraging experimentation and working closely with professionals in the field will help discover innovative ways to utilize this technology. The insights gained from these sessions will guide future experiments and inform the development of practical guidance for the public health community.

AI Roadmap for Local Health Departments

Building on the Community of Practice and incorporating participant feedback, an AI Roadmap was developed for LHDs to guide future efforts over the coming years. This roadmap outlines key goals and implementation strategies for integrating AI into local public health services.

The implementation will focus on two main areas: increasing familiarity with generative AI, including training on effective prompting techniques, and exploring the use of AI models to automate tasks and find programmatic solutions for specific public health functions, such as disease surveillance. This roadmap aims to start building a path from current generative AI applications to more advanced models and systems, supported by comprehensive training and policy development. The roadmap will continue to evolve with new insights and feedback from health departments, ensuring it remains a dynamic tool for advancing public health through AI.

Generative AI has already proven valuable in various Kansas LHD applications. In the "AI in Public Health" Community of Practice, two specific examples were highlighted. One county used ChatGPT to quickly develop procedures for tasks routinely missed by staff, which were then refined and implemented. Another health department used AI to transform a brief event description into an engaging newspaper article and multiple social media posts, significantly reducing the time required and improving timely outreach.

Kansas is Part of the National AI Policy Project

The Kansas Health Institute, Health Resources in Action and Wichita State University's Community Engagement Institute have launched a project to develop an AI policy or guidelines template for public health organizations, with the goal of making it available by 2025. The project includes a literature review of AI research and an analysis of existing AI-related guidelines and policies at both the city and state levels. The team is specifically focusing on identifying approaches to address data privacy, the impact on vulnerable populations, transparency, bias and equity within the AI guidelines or policy template. This initiative is supported by the Centers for Disease Control and Prevention's Public Health Infrastructure Grant.

Next Steps

What are the next steps for ethically advancing AI in public health in Kansas? The following steps, which have been implemented in other localities, could be considered by the public health community in Kansas. The list is not comprehensive or presented in hierarchical order.

• Conduct an inventory of AI systems or systems with AI components currently in use. This will help identify training needs and ensure transparency for stakeholders regarding operations.
• Identify how staff and vendors use AI, including the utilization of generative AI.
• Develop a framework for reviewing AI if there is an interest in procuring an AI tool.
• Foster cross-sector collaboration with AI developers.
• Create guidelines for public engagement with AI.
• Explore opportunities to bring AI expertise to governmental public health.
• Develop AI guidance and policies.
• Develop an AI steering committee at the state, county or city levels.

AAron Davis is the Director of the Center for Public Health Initiatives at the Wichita State University Community Engagement Institute.