{"id":"W4396573797","doi":"10.1093/ced/llae158","title":"Assessing GPT-4’s diagnostic accuracy with darker skin tones: underperformance and implications","year":2024,"lang":"en","type":"letter","venue":"Clinical and Experimental Dermatology","topic":"Artificial Intelligence in Healthcare and Education","field":"Medicine","cited_by":4,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto","funders":"","keywords":"Medicine; Dermatology; Audiology","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":[0.0001074585,0.0002480939,0.0005428457,0.00009400236,0.0001807537,0.0001119327,0.00007213555,0.0006609212,0.00006345411],"category_scores_gemma":[0.0002433074,0.0001849544,0.0000696381,0.0001055963,0.0007325532,0.0001775305,0.00008411164,0.001200215,0.0001003064],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000381549,"about_ca_system_score_gemma":0.0002369105,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00008923698,"about_ca_topic_score_gemma":0.000008221718,"domain_scores_codex":[0.9981826,0.00009456821,0.0007253749,0.0005644088,0.0001113988,0.0003216791],"domain_scores_gemma":[0.9972953,0.002056642,0.0001338636,0.0002847607,0.0000639241,0.0001654958],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0001101212,0.0003094242,0.1051229,0.001285599,0.00021594,0.0007090155,0.001409791,2.330434e-7,0.00005988761,0.0003146195,0.8500024,0.04046001],"study_design_scores_gemma":[0.0004913156,0.001050889,0.03800635,0.002025491,0.000653426,0.01345377,0.005620429,0.000426978,0.004979866,0.003117534,0.9291911,0.000982871],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"commentary","genre_gemma":"empirical","genre_scores_codex":[0.4676946,0.006575285,0.00005941396,0.5239155,0.0006967835,0.0004052197,0.000006130281,0.00005687405,0.0005902544],"genre_scores_gemma":[0.7312999,0.001276678,0.0002732052,0.2654129,0.001226378,0.0001380746,0.0001714316,0.00003502769,0.0001664658],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.2636053,"threshold_uncertainty_score":0.7542223,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1912669036070457,"score_gpt":0.5078072894596163,"score_spread":0.3165403858525706,"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."}}