Factors influencing people’s emotional experiences when using ChatGPT for health information: a cross-sectional web-based questionnaire survey in the UK

Understanding how individuals emotionally respond to AI-driven chatbots in healthcare is crucial for optimising user experience and ensuring effective health information delivery. This study aimed to investigate the factors influencing people’s emotional experiences when using ChatGPT for health information. A cross-sectional, web-based questionnaire survey was conducted in August 2024, targeting a demographically representative sample in the UK. A total of 462 respondents participated. Data was analysed descriptively and using a multiple stepwise linear regression analysis to explore the relationship between people’s emotional responses, measured using the PANAS tool, and user characteristics. The independent variables included (1) sociodemographics, (2) general (non-health-related) information-seeking behaviours towards the use of ChatGPT, and (3) health-related information-seeking behaviours using ChatGPT (including perceived aesthetics, usability, and level of trust towards ChatGPT). The results showed that older users (35+ years) and ethnic minorities (Asian or Black participants) reported higher positive emotions, while younger users (25–34 years) experienced more negative emotions when using ChatGPT for health information. Positive emotions were strongly associated with confidence in using ChatGPT, perceived trust, and perceived aesthetics, while negative emotions were linked to longer conversations, lower perceived trust, and poor functional quality. These findings highlighted the need to improve ChatGPT’s aesthetics, usability, trustworthiness, and integration to enhance users’ positive emotions in health information-seeking contexts.<p></p>