Integrating large language models into medical education: A commentary on opportunities, challenges, and future directions
Bibliographic record
Abstract
Shortly following its release in November 2022, OpenAI’s ChatGPT gained notoriety in the medical education community for its ability to perform at or near the passing threshold on the United States Medical Licensing Examination (USMLE). Although there is an overwhelming amount of excitement surrounding artificial intelligence (AI) and its potential to revolutionize medical training, this commentary seeks to explore both the opportunities and challenges posed by incorporating large language models (LLMs) such as ChatGPT into medical education. To evaluate ChatGPT’s impact in the context of problem-based learning (PBL) medical education, we conducted two initial studies. The first study assessed ChatGPT’s performance on concept application exercises (CAEs)—short-answer assessments used in our program to gauge student progress. After establishing ChatGPT’s performance on CAEs, our second study aimed to evaluate ChatGPT’s ability to effectively grade student-generated responses. Our results reveal that ChatGPT not only outperforms students who are marginally passing but also grades these assessments with promising alignment to human grading practices. Our team’s future research plans include examining ChatGPT's ability to provide effective feedback, generate discerning assessment questions, create realistic training scenarios, and support continuous professional development. Although we are optimistic about future applications of LLMs, we emphasize the need for an AI-assisted approach that employs human oversight to mitigate the inherent risks associated with LLMs, such as bias perpetuation, inaccuracies, over-reliance, and potential misuse. Ultimately, through the thoughtful and evidence-based implementation of these new tools, we believe AI can be harnessed to augment rather than undermine the quality and effectiveness of medical education.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".