As we’ve previously shared, Artificial Intelligence (AI), specifically generative AI, provides new opportunities to advance the health and well-being of people and populations through better and more efficient communication. Our work in this area over the past year has led to a successful collaboration with a group of public health and technology leaders in developing guidelines around the ethical use of AI in public health communication. We believe these guidelines will help harness the enormous power of technologies like OpenAI, Anthropic’s Claude, and Google Gemini.
Our aim was to create something that could help public health communicators, who are considering incorporating AI tools into their work, keep their mission to the public’s wellbeing in mind while also being practical and transparent.
This was difficult work; especially given the speed at which AI developments continue to evolve, the concerns that have been raised about the quality of AI results, and the industry’s critical need to keep pace with this growth.
What we cover in our summary and working paper are five principles that we believe should guide the usage of AI in public health communications:
Respectfully Protect the Public
An important goal of public health communication is to safeguard the public’s health and wellbeing. When using AI technologies, two aspects need to be kept in balance: (1) protective practices around personal information and (2) transparent explanations about AI usage.
Build Credibility through Accuracy
Ensuring accuracy in health information is paramount. Public health organizations bear the ultimate responsibility for providing guidance that is not only timely and complete but, above all, accurate and trustworthy.
Center Humans and Their Judgment
In public health communication, AI should serve as an assistant to human judgment, not a replacement. Humans should be the ultimate judges, evaluators, and authors of any message, ensuring that the creative works of others are appropriately referenced.
Advance the Health of All Communities
AI tools must be used by public health communicators to provide opportunities for all communities to achieve their best possible health. In particular, communications must strive to promote fairness and justice, and to avoid unintentional harm.
Iterate Responsibly
Given the rapid development of AI, its use must involve departments continuously “iterating,” or testing and trying out AI technology again and again, in ways that are responsible.
You can download a summary with these guidelines here: https://www.discourselabs.org/ai-ethics-public-health. There, you can also find a full working paper that explains some of the research behind our thinking. It includes a discussion of opportunities and risks related to AI usage. It also contains a checklist, policy template, and examples of disclosure.
One thing that became abundantly clear during the past year is how much work is left to be done in this area, especially since AI continues to develop so rapidly. Developments in technology, regulations, and research will require changes to guidance, which is why we are only sharing a “first edition” starting point. The paper includes suggestions for future development, with the expectation of ongoing progress in this area.
We extend our sincere appreciation to those who contributed to the successful completion of this significant work. We are particularly thankful for the dedicated efforts of Andrea Bagnall Degos, Ann Rowe, Elizabeth Perez, Glen Nowak, Johnnie (Chip) Allen, Khalilah LeGrand, Kristin Howard, Mark Miller, Michael Grela, Amelia Burke-Garcia, Amanda Yarnell, and Elizabeth (Betsy) Mitchell. Their commitment and countless hours spent evaluating the benefits and implications of using AI in public health communications were invaluable. In addition, we want to thank those who provided feedback that came to us through open channels; all of it helped to make this first version of the guidelines even better. Finally, we also are grateful to Brooke Wadsworth and Elizabeth Harris who meticulously supported the working group throughout the process.
Of course this is just the beginning, and we look forward to the future. We are hopeful that public health professionals will share these guidelines widely and integrate them into their AI initiatives. To improve this work and inform future editions, we also seek your ongoing feedback as your organizations progress in this new era.
–––––––
Connie Moon Sehat is Director of Research at Discourse Labs and Robert Jennings is Executive Director of the National Public Health Information Coalition.
