Shaping the future: A pilot study on how AI-powered chatbots shape patient perceptions of pharmacist roles

To investigate how AI-powered chatbots influence patient perceptions of pharmacist roles compared to traditional educational methods and evaluate whether such technology can address knowledge gaps in pharmacy services.

A randomized controlled trial using a between-subject design was conducted through a crowdsourcing platform with 280 participants: baseline control (n = 82), educational video (n = 101), and AI chatbot interaction (n = 97). The survey assessed factors such as trust, comfort, and communication regarding pharmacist services using 7-point Likert scales. The AI chatbot utilized ChatGPT4-o, programmed with identical information to the video content to ensure experimental control. Participants in the chatbot group submitted an average of 3.03 prompts (range: 0-25). Statistical analysis included t-tests, ANOVA, and regression analysis of engagement metrics including word count and question types.

Significant differences emerged across treatment groups in willingness to seek pharmacist information, comfort asking questions, and awareness of medication monitoring. Participants who actively engaged with the AI chatbot showed significantly improved perceptions of pharmacist roles. Sentiment analysis revealed predominantly neutral (n = 252) or positive (n = 95) interactions, with few negative engagements. Participants preferred detailed "wh" questions (n = 170) over yes/no questions (n = 103), with frequent inquiries about fundamental pharmacist roles. Regression analyses confirmed that engagement depth positively correlated with improved perceptions, suggesting deeper interaction led to better understanding of pharmacist capabilities beyond basic dispensing functions.

Although both video and AI chatbot interventions improved patient knowledge, engagement level-not modality type-emerged as the critical determinant of outcomes. AI chatbots can effectively address knowledge gaps about pharmacist roles when patients actively participate in the learning process.

AI-powered chatbots can serve as valuable supplements to traditional patient education, particularly addressing limited understanding of pharmacist capabilities. Implementation should focus on interfaces that encourage active participation, as engagement emerged as the key determinant of educational effectiveness while complementing rather than replacing human pharmacist interactions.