Skip to main content
RACS ASC 2026
A systematic review of Large Language Model’s capabilities in General Surgery educational content creation and simulation
Poster

Poster

Disciplines

General Surgery

Presentation Description

Institution: Wollongong Hospital - NSW, Australia

Generative artificial intelligence, particularly Large Language Models (LLMs), are increasingly integrated into surgical education. While early studies have focused on medical student examinations, its role in postgraduate general surgery training remains unclear. This study reviews the educational utility of LLMs as tools in general surgery training. A systematic search of PubMed and Scopus was conducted without date restriction. Eligible studies evaluated generative AI models within general surgery education, training, or recruitment. Included populations comprised of medical students, junior doctors, general surgery trainees, faculty as examiners, or AI models as learners. Two reviewers independently screened studies, extracted data, and assessed study quality and risk of bias. This study was conducted as part of a larger review on LLM use across surgical exams, education and workforce recruitment. Thirty-four articles met inclusion criteria, with nine studies directly evaluating AI as an educational adjunct in general surgery. Four studies used LLMs to create surgical case-based scenarios that promoted reflective learning and communication. One study found a significant improvement in correct answers after seeking resources on ChatGPT and rated its usefulness as 6.6/10. Two studies found that LLMs were able to provide feedback comparable to that of experienced surgeons and were considered useful for skill improvements. Limitations identified included "hallucinations" (factual inaccuracies), variable clinical reasoning in complex scenarios, and overconfident incorrect responses. LLMs demonstrate significant promise as adjunctive tools in general surgery education, particularly for simulation-based learning and automated feedback. These models offer a scalable solution to enhance training equity, providing high-quality educational support in regional settings where faculty access may be limited. However, given the risk of inaccuracies, current evidence supports LLMs as a supplement rather than a replacement for traditional teaching.
Presenters
Authors
Authors

Dr Mingchun Liu - , Dr Nicholas Shannon -