Answers are grounded in this paper and cite their sources.
Randomized trial compares AI-assisted learning and traditional methods in emergency care training, suggesting differences in skill acquisition.
Background As the integration of artificial intelligence (AI) into health professions education becomes increasingly inevitable, educators face the challenge of determining how to use AI tools effectively and safely. Despite the potential of AI-assisted self-directed learning, limited evidence exists comparing its impact to traditional methods, especially regarding psychomotor and critical thinking skills in emergency care training. Methods This pilot randomized controlled trial compared AI-assisted self-directed learning, lecture-based instruction, and simulation-based education in teaching emergency choking management to pre-professional health students. Twenty students (n = 6-7 per group) enrolled in a two-week preparatory program were randomly assigned to one of three instructional groups: lecture, simulation, or AI-assisted self-directed learning using ChatGPT-4o. Notably, the lecture and simulation groups received roughly 60-minute faculty-led sessions, while the AI-assisted group received the standardized educational outline and was asked to self-study using ChatGPT-4o. All participants completed a post-intervention simulation scenario and a knowledge assessment on the third day of the two-week program (multiple-choice and essay), followed by a repeat assessment on the final day of the program (nine days later). Simulation performance was scored using an expert-developed checklist. Essay responses and AI interaction transcripts were evaluated using a modified Holistic Critical Thinking Scoring Rubric. A focus group was conducted to explore learner experiences. Results Twenty pre-professional students participated (lecture n = 7, simulation n = 6, AI-assisted n = 7). Students in the lecture and simulation groups demonstrated higher performance and confidence levels than those in the AI-assisted group, with statistically significant differences on performance and observer-rated confidence (inter-rater reliability α = .83). Analysis of AI transcripts revealed substantial variability in prompting strategies, metacognitive engagement, and depth of learning. Focus group data indicated both perceived benefits and psychological barriers to AI use, including feelings of uncertainty and guilt. Conclusion In this small pilot, AI-assisted self-directed learning appeared to underperform faculty-supported lecture and simulation formats in teaching emergency care skills. Because the AI group received less instructional support than the lecture and simulation groups, these findings should be interpreted as a comparison of AI as a self-directed modality against faculty-led instruction, rather than AI-assisted teaching. The data are suggestive that early development of AI literacy, prompting competency, and psychological safety may be important for successful integration. Future research should involve larger, more diverse cohorts, baseline assessment, comparable instructional support across arms, and explicit study of AI literacy and prompt-design scaffolding for blended instructional models.
No takes yet. Share an insight, caveat, or question.
Palaganas et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: