Artificial Intelligence and Large Language Models: A Case-Based, Peer-Teaching Workshop for Preclinical Medical Students

Clinical Snapshot

Key Findings

• P Value: P < .01 for all six knowledge/self-efficacy items; mixed (direction and significance unspecified) for five attitude/behavioural intent items

• Effect Size: Not reported; mean pre- and post-scores and standard deviations are not provided in the abstract

• Primary Outcome: Statistically significant increase in mean ratings across six 5-point Likert scale items assessing AI-related knowledge and self-efficacy following the combined lecture and workshop session

• Nnt Or Sensitivity: Not applicable for this educational intervention study; no NNT, sensitivity, specificity, or hazard ratio reported. Kirkpatrick Level 1 (reaction) and Level 2 (learning — self-reported) outcomes only.

• Confidence Interval: Not reported

Clinical Application

The session format (60-minute lecture + 100-minute workshop) is feasible within standard medical school timetabling. The peer-teaching component reduces faculty resource burden. Case materials published via MedEdPORTAL are freely accessible. However, local adaptation would be required to contextualise AI tools, regulatory frameworks, and clinical examples to specific national or institutional settings. Faculty with AI literacy expertise are needed for quality oversight. Australian medical schools are increasingly confronted with student use of LLMs such as ChatGPT and Google Gemini in study and clinical reasoning. The Medical Board of Australia and AHPRA have not yet issued specific guidance on AI use in medical training, though the TGA has begun regulating AI-based Software as a Medical Device (SaMD). The RACGP and other specialist colleges have acknowledged AI's growing role in clinical practice. Australian medical education bodies (AMC, MDANZ) would benefit from developing standardised AI literacy frameworks. This curriculum resource could serve as a template for Australian schools, though content would need to reflect Australian clinical guidelines, medicolegal frameworks, and the specific AI tools available in the Australian healthcare ecosystem. No PBS or TGA-specific AI therapeutics are directly implicated in this educational study. Preclinical (preclerkship) medical and dental students in Years 1–2 of undergraduate medical education programs. Potentially applicable to allied health, nursing, and pharmacy students with curriculum adaptation.

Abstract

INTRODUCTION: Artificial intelligence tools have rapidly become integrated in health care settings and are quickly affecting medical student education. However, there remains limited formal teaching on such tools in student curricula. This project sought to introduce preclerkship medical students to the basics of large language models and allow them to practice ways to best use these tools to supplement their learning. METHODS: The authors designed, implemented, and evaluated a 60-minute lecture and 100-minute workshop for second-year medical students. The workshop included interactive cases covering various aspects of artificial intelligence use, and some groups were led entirely by student leaders, allowing for peer teaching. RESULTS: One hundred sixty-eight students from Harvard Medical School and Harvard School of Dental Medicine were enrolled in this session. Anonymous pre- and postsession surveys (N = 124 and N = 62, respectively) were collected and compared via unpaired t test assuming unequal variance and showed statistically significant increase in the mean ratings of six 5-point Likert scale questions assessing artificial intelligence-related knowledge/self-efficacy (P < .01) and mixed changes to 5 questions relating to attitudes/behavioral intent. DISCUSSION: Our artificial intelligence and large language model session provides a framework for teaching medical students, early in their educational journey, the basics of these tools. Our session provides interactive exercises to illustrate how best to leverage such tools while also discussing their potential risks. Such education will be important to incorporate into medical student curricula as artificial intelligence technologies grow increasingly common.

References

1. 1.Choy, B. C., Shanmugam, H., Redden, H. W., Duran, E., Prince, J., Lee, R. S., Chan, J., Limbania, D., Berwick, J., & Rodman, A. (2026). Artificial intelligence and large language models: A case-based, peer-teaching workshop for preclinical medical students. MedEdPORTAL: The Journal of Teaching and Learning Resources. https://doi.org/10.15766/mep_2374-8265.11621