Health & Medicinearticle2026-08-29

Knowledge, use and perceptions of AI-based chatbots among medical undergraduates in Sri Lanka: a cross-sectional study

Open access0 citations

Abstract

AI-based chatbots are increasingly used in medical education, but evidence from Sri Lanka is limited. This study assessed medical undergraduates’ knowledge, use, perceptions, barriers, and expectations regarding these tools. A cross-sectional online survey included 400 pre-clinical and clinical medical undergraduates. The locally piloted questionnaire had a Cronbach’s alpha of 0.856. Exploratory univariable binary and ordinal logistic regression produced unadjusted odds ratios (ORs). The proportional-odds assumption was tested, and common ORs were withheld when it was violated. Among respondents, 65.8% were female; 98.5% were aware of AI chatbots and 98.5% had used them for learning, with peers the main awareness source (58%). Chatbots were rated effective or very effective by 67.5%, trustworthy by 79.3%, and as improving learning experience by 85.8%. Major benefits were time saving (81.7%), constant availability (81.5%), and personalised learning (71.7%); concerns included inaccurate information (76.2%), reduced critical thinking (66.9%), and overdependence (49.6%). In models satisfying the proportional-odds assumption, clinical students reported greater perceived effectiveness (OR 1.88, 95% CI 1.29–2.73, p = 0.001), better experience (OR 1.97, 1.35–2.88, p < 0.001), greater accessibility (OR 2.40, 1.66–3.47, p < 0.001), greater future importance of AI proficiency (OR 1.77, 1.23–2.55, p = 0.002), and greater ethical inappropriateness (OR 1.86, 1.24–2.80, p = 0.003). Male students reported greater knowledge (OR 2.26, 1.53–3.36, p < 0.001), more frequent use (OR 1.54, 1.00–2.37, p = 0.049), greater accessibility (OR 1.62, 1.10–2.38, p = 0.014), and greater ethical inappropriateness (OR 1.91, 1.25–2.93, p = 0.003). AI chatbots were widely used and generally perceived favourably, alongside concerns about accuracy, ethics, and dependence. Structured AI literacy, ethical guidance, and faculty-led curricular integration are needed.

// Source

View paper (DOI)Open access versionOpenAlexBMC Medical EducationPublished 2026-08-29

Authors: Rakitha Higgoda, Anuki Alahakoon, J. Indrakumar

Institutions: University of Sri Jayewardenepura