THURS-091 - Testing the Feasibility of Generative AI for Campus Health Messaging

Brief Abstract Summary: Explore the feasibility of using generative AI to create effective health messages about nicotine pouch risks for campus health programs. Generative AI offers an opportunity to quickly create tailored content at a low cost. There is little evidence on campus programs using generative AI. This project explores the feasibility of using generative AI for campus health educators. This project will use two phases: 1) a focus group with health educators assessing feasibility, and 2) intercept interviews with college students. Health educators will be provided with instructions for how to prompt AI programs to create 4 tobacco messages followed by questions of experience. Intercept interviews will compare likability between human and AI-made messages. A guide on effectively using the AI programs will be made and shared with the campus program. Findings will provide evidence of how AI can be integrated in health messaging.

Detailed abstract description: Nicotine pouches, such as Zyn, are a newer tobacco product that is most popular among college-aged students. Many college students hold misconceptions about the harm of nicotine pouches and often underestimate or are unaware of the health risks. A trusted source for delivering timely, accurate, and relevant health information are university health programs that can help increase this awareness among college students. Yet, campus health educators often face limited time and resources for developing effective health messaging campaigns. Generative Artificial Intelligence (AI), such as ChatGPT and Canva Pro, offers the opportunity to quickly create engaging and tailored health promotion content at a low cost. However, there is little evidence on the accuracy and feasibility of campus health wellness programs using generative AI for health messaging. This project focuses on exploring the feasibility and efficacy of using generative AI for campus health messaging. This project will use two phases including (1) a focus group with student peer health educators to assess the feasibility of using generative AI for developing nicotine pouch cessation messages, and (2) intercept interviews with college students to gather initial feedback on AI generative messages and compare them to human made messages. Working with a campus health wellness program in Central California, a group of health educators will be provided with instructions for how to prompt ChatGPT and Canvas Pro to create 4 of tobacco cessation messages. After that, a focus group with the health educators will be conducted to explore their feedback on the prompts in terms of ease of use, how much time it took, and how to be used in their current work. Following, intercept interviews with college students will be used to compare perceived likability between human-made and generative-AI messages developed during the focus groups. Finally, a guide on prompts and strategies for effectively using Canvas Pro and ChatGPT will be developed based on the findings and shared with the campus program. These findings will provide more evidence how AI can be integrated in messaging for campus wellness programs and provide a foundation for more long-term evaluation of the effectiveness of AI-generated health messaging for promoting health behaviors among college students.