Shared Values, Shared Responsibility, and Overwhelmingly Positive: Investigating a Targeted COVID-19 Public Health Messaging Approach Using Generative Artificial Intelligence

Misinformation during the COVID-19 pandemic was linked to noncompliance. Nonadherence, such as vaccine hesitancy, in a health emergency can lead to costly and deadly consequences. In an environment of mistrust and misinformation, quality, effective information that engages the public, shapes attitudes, and encourages behavior change is crucial. However, communicating effectively and ensuring compliance during the pandemic was challenging for health communicators, resulting in burnout and attrition. Generative artificial intelligence (AI), therefore, may be useful for communicators designing targeted health messages, particularly for vulnerable populations that are prone to misinformation. In this study, the authors explored the use of ChatGPT 3.5, a generative AI software, in the design of targeted COVID-19 booster vaccine messages for a variety of demographic groups (e.g. political leaning, religion, and conspiracy theorists). Sentiment and inductive thematic analyses revealed variations in targeted group messaging that leaned heavily on positive valence and situated adherence to public health guidelines within narratives about individuals' roles and responsibilities to protect self, community, and values. The authors posit that an audience-centered, positive framing messaging approach that appeals to individuals' shared values and responsibilities may be effective in priming audiences and shaping attitudes to minimize the impact of misinformation.