{"id":"W4386893393","doi":"10.1016/j.ajp.2023.103770","title":"Appraising the performance of ChatGPT in psychiatry using 100 clinical case vignettes","year":2023,"lang":"en","type":"article","venue":"Asian Journal of Psychiatry","topic":"Artificial Intelligence in Healthcare and Education","field":"Medicine","cited_by":77,"is_retracted":false,"has_abstract":false,"ca_institutions":"Queen's University","funders":"","keywords":"Grading (engineering); Mental health; Psychiatry; Psychology; Medicine; Medical education","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":[0.001908385,0.0001078534,0.0003243092,0.0003269476,0.0001199591,0.00001272824,0.000137923,0.0001207519,0.00003656316],"category_scores_gemma":[0.0001701405,0.00007645923,0.0001866948,0.0007914932,0.0001301166,0.0001447213,0.00001657177,0.0006143798,0.00001627607],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003312861,"about_ca_system_score_gemma":0.001049342,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00007498373,"about_ca_topic_score_gemma":0.0001419168,"domain_scores_codex":[0.9978611,0.0001401958,0.001376793,0.0001291806,0.0002538497,0.0002388319],"domain_scores_gemma":[0.9987192,0.0001276747,0.0005989064,0.0002484269,0.0001743266,0.0001314789],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0002017658,0.0002195202,0.9536943,0.0001975298,0.00004444466,0.00008413773,0.001730057,0.0001553152,0.00006364472,0.0002566151,0.002154895,0.04119778],"study_design_scores_gemma":[0.001171365,0.004667385,0.8224229,0.006030607,0.000521987,0.02642597,0.1118139,0.01609376,0.0008770164,0.00826311,0.001184829,0.0005271597],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9642723,0.000911107,0.00007681487,0.02727043,0.007095254,0.0001448444,9.646657e-7,0.000009522693,0.0002187361],"genre_scores_gemma":[0.9926578,0.0001624083,0.004257001,0.0005279781,0.002356583,0.000001127952,0.000001107699,0.00001797764,0.00001799771],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1312714,"threshold_uncertainty_score":0.3117917,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1655565224239388,"score_gpt":0.4824548562779592,"score_spread":0.3168983338540204,"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."}}