{"id":"W4417534853","doi":"10.1371/journal.pdig.0001156","title":"Evaluating ChatGPT-4 in the development of family medicine residency examinations","year":2025,"lang":"en","type":"article","venue":"PLOS Digital Health","topic":"Artificial Intelligence in Healthcare and Education","field":"Medicine","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"St. Michael's Hospital; University of Toronto; University Health Network","funders":"","keywords":"Quality (philosophy); Educational measurement; MEDLINE; Medical school; Residency training","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001131956,0.00007932405,0.0002203096,0.0002413968,0.0001289589,0.00001071543,0.0001021398,0.00004061282,0.00001696195],"category_scores_gemma":[0.00126826,0.00005505398,0.00001997166,0.0006354206,0.00006693134,0.00009996095,0.00001781937,0.0001585191,0.00001386352],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002194907,"about_ca_system_score_gemma":0.002321549,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0003958722,"about_ca_topic_score_gemma":0.0003593664,"domain_scores_codex":[0.998346,0.00007269839,0.0008328398,0.0001633581,0.0003727823,0.0002122855],"domain_scores_gemma":[0.9988169,0.0005662169,0.0001319023,0.000221492,0.0002036418,0.00005980979],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"observational","study_design_scores_codex":[0.00006188131,0.0007217802,0.05706728,0.00132534,0.00002227787,0.000001633378,0.05839785,0.000003741558,0.0003437914,0.0061873,0.001252424,0.8746147],"study_design_scores_gemma":[0.0003622165,0.001637295,0.8536232,0.006663478,0.00003895964,0.000007592376,0.1089832,0.001174028,0.003630267,0.02170875,0.001976151,0.0001948315],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9569113,0.0009212406,0.0002653291,0.02169116,0.0001463816,0.0007481978,0.000002145682,0.00001956849,0.01929466],"genre_scores_gemma":[0.996795,0.00005679583,0.0009256477,0.001751197,0.00007186483,0.00006389844,0.00003619108,0.000005507908,0.0002939285],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.8744199,"threshold_uncertainty_score":0.4118328,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.4848804161110564,"score_gpt":0.5470551834505694,"score_spread":0.06217476733951299,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}