Using AI chatbots (e.g., CHATGPT) in seeking health-related information online: The case of a common ailment

Abstract

In the age of AI, healthcare practices and patient-provider communications can be significantly transformed via AI-based tools and systems that distribute Intelligence on the Internet. This study employs a quantitative approach to explore the public value perceptions of using conversational AI (e.g., CHATGPT) to find health-related information online under non-emergency conditions related to a common ailment. Using structural equation modeling on survey data collected from 231 respondents in the US, our study examines the hypotheses linking hedonic and utilitarian values, user satisfaction, willingness to reuse conversational AI, and intentions to take recommended actions. The results show that both hedonic and utilitarian values strongly influence users' satisfaction with conversational AI. The utilitarian values of ease of use, accuracy, relevance, completeness, timeliness, clarity, variety, timesaving, cost-effectiveness, and privacy concern, and the hedonic values of emotional impact and user engagement are significant predictors of satisfaction with conversational AI. Moreover, satisfaction directly influences users' continued intention to use and their willingness to adopt generated results and medical advice. Also, the mediating effect of satisfaction is crucial as it helps to understand the underlying mechanisms of the relationship between value perceptions and desired use behavior. The study emphasizes considering not only the instrumental benefits but also the enjoyment derived from interacting with conversational AI for healthcare purposes. We believe that this study offers valuable theoretical and practical implications for stakeholders interested in advancing the application of AI chatbots for health information provision. Our study provides insights into AI research by explaining the multidimensional nature of public value grounded in functional and emotional gratification. The practical contributions of this study can be useful for developers and designers of conversational AI, as they can focus on improving the design features of AI chatbots to meet users’ expectations, preferences, and satisfaction and promote their adoption and continued use.