Answers are grounded in this paper and cite their sources.
The integration of generative artificial intelligence (AI) systems into medical education carries the risk that persuasive yet inaccurate information may be accepted without critical evaluation. This study aimed to examine the relationships between medical students’ general trust in AI, their subjective appraisal of the reliability of AI-generated outputs, analytical thinking skills, information verification behaviors, and their overall anatomical classification accuracy when evaluating AI-generated anatomical content. This cross-sectional observational study included 82 s-year medical students. Analytical thinking levels were assessed using the Cognitive Reflection Test (CRT), and general reliability perceptions toward AI were evaluated using an 8-item AI Reliability Perception Scale, whose internal consistency was assessed in the present sample. Students were presented with a 10-statement text described as generated by an AI assistant and containing deliberately embedded anatomical errors in 40% of the statements. Without prior information about the presence of errors, students evaluated each statement as true or false. Overall anatomical classification accuracy was scored according to an expert-determined answer key. Self-reported verification and questioning tendencies were assessed using a 3-item rubric based on the lateral reading approach. Data were analyzed using multivariable linear regression to examine associations between the study variables and overall anatomical classification accuracy. A negative correlation was observed between general trust in AI scores and overall anatomical classification accuracy ( r = -0.840, p < 0.001). Similarly, students who stated that AI-generated outputs were generally “reliable” demonstrated significantly lower overall anatomical classification accuracy (β = -0.924, p = 0.049). These findings suggest that both general trust in AI and subjective appraisals of AI-generated output reliability were associated with poorer overall anatomical classification accuracy. Analytical thinking tendency (CRT; β = 0.957, p < 0.001) and a high grade point average (GPA of 3.00–3.49: β = 1.759, p < 0.001; GPA of 3.50–4.00: β = 2.300, p < 0.001) were identified as independent positive predictors of performance. This study identified associations between medical students’ performance in evaluating AI-generated anatomical information and cognitive reflection, self-reported verification tendencies, and perceptions of AI reliability. These findings highlight the potential relevance of critical evaluation skills in AI-supported anatomy education. However, the cross-sectional design does not establish causal relationships or the cognitive mechanisms underlying these associations. Further longitudinal and experimental studies are needed to determine whether educational interventions targeting analytical reasoning and information verification can improve students’ evaluation accuracy.
No takes yet. Share an insight, caveat, or question.
Şenol et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: