AI- and VR-supported learning configurations linked to higher feedback-literacy scores in randomized medical-student trial

A randomized controlled study of 68 fourth-year medical students found that two multicomponent interventions combining artificial intelligence-supported reflective activities with virtual-reality clinical scenarios were associated with higher supervisor-rated feedback-literacy scores at post-test than usual clinical clerkship activities. The summary does not identify which individual intervention component was responsible or establish a specific causal mechanism.

Students were assigned to an AI–VR configuration, an AI–VR configuration that also included peer dialogue, or a control condition during their transition to clinical clerkships. Both intervention groups completed a preparatory workshop and mandatory AI-supported reflection within VR-based clinical scenarios. The configurations differed in the inclusion of peer dialogue and in the distribution of active VR participation. Supervisors assessed feedback-literacy behaviors before and after the study using the study-specific Supervisor Observational Checklist for Feedback Literacy, or SOC-FL. Quantitative comparisons of post-test scores across the three groups were adjusted for baseline SOC-FL scores.

After adjustment for baseline scores, the AI–VR group had an adjusted post-test score of 21.12 points, with a 95% confidence interval of 18.19 to 24.05, while the AI–VR-plus-peer group had an adjusted score of 23.45 points, with a 95% confidence interval of 20.59 to 26.31. Both comparisons with the control group had p-values below .001. The two intervention configurations did not differ significantly from one another: the reported difference was 2.33 points, with a 95% confidence interval of −0.53 to 5.19 and p = .149. Because peer dialogue was not the only feature distinguishing the configurations, the results do not show that adding dialogue produced the numerical difference between them.

The abstract also reports a thematic analysis of 36 student reflections, without specifying whether these came from 36 distinct participants. Students in the AI–VR-plus-peer group more frequently described collaborative sense-making, normalization of uncertainty and reduced emotional threat. Those in the AI–VR group reported varied experiences of independent AI use and individual reflection, while control-group reflections tended to characterize uncertainty as evidence of incompetence and AI as a confirmation tool. These qualitative accounts describe reported experiences rather than measured effects on learning outcomes.

This account is based on the published abstract rather than a full assessment of the study and its detailed protocol. The abstract does not report group sizes, allocation procedures, attrition, assessor masking, the characteristics of the SOC-FL scale or longer-term outcomes. The population was limited to 68 fourth-year medical students at the transition into clinical clerkships. Although random assignment was used, the reported results do not demonstrate a specific causal mechanism or broader clinical benefit.

Subject of Research: Randomized trial of AI- and VR-supported learning interventions and supervisor-rated feedback literacy in fourth-year medical students

Article Title: Effects of AI- and VR-supported multicomponent learning intervention configurations on supervisor-rated feedback-literacy behaviors in medical students: a randomized controlled trial

Article References: Sadegh, T., Mohammadi, M., Jahani, J., & Salimi, G. (2026). Effects of AI- and VR-supported multicomponent learning intervention configurations on supervisor-rated feedback-literacy behaviors in medical students: a randomized controlled trial. BMC Medical Education. https://doi.org/10.1186/s12909-026-10500-7

Image Credits: AI Generated

DOI: 10.1186/s12909-026-10500-7

Keywords: medical education, feedback literacy, artificial intelligence, virtual reality, medical students, randomized controlled trial, clinical clerkships, reflective learning, peer dialogue, Effects, VR-supported, multicomponent