Direct comparison of GPT-4 and human physicians in MKSAP-19 multiple-choice questions
Bibliographic record
Abstract
Several studies have compared scores of artificial intelligence (AI) algorithms on medical multiple-choice questions (MCQs) with reference standards. In this study, the authors directly compared scores of an AI algorithm (Generative Pre-trained Transformer 4 [GPT-4]) with those of clinicians. A stratified random sample of 600 Medical Knowledge Self-Assessment Program-19 (MKSAP) MCQs were inputted into GPT-4. The proportion of questions answered correctly was compared with the answer selected by the majority of the MKSAP clinician testing group (consensus clinicians) and the proportion of the MKSAP clinician testing group who selected the correct answer (average clinician). GPT-4 answered 496 questions correctly (82.7%, 95% CI 79.6 to 85.7). This was significantly less than the consensus clinician (88.0%, 95% CI 85.4 to 90.6; McNemar's T statistic 10.0, P = 0.0015) but was significantly greater than the average clinician (64.7%, 95% CI 63.1 to 66.3; paired T statistic = 12.7, P < .0001). Results did not significantly vary by specialty. GPT-4 scored significantly lower than the consensus clinicians, but significantly greater than the average clinician, on MKSAP MCQs.
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 machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.023 | 0.125 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.003 | 0.002 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 0.001 |
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 source (direct Gemma or distilled Codex), 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".