{"id":"W4385365799","doi":"10.1016/j.jacr.2023.06.025","title":"“Shortcuts” Causing Bias in Radiology Artificial Intelligence: Causes, Evaluation, and Mitigation","year":2023,"lang":"en","type":"review","venue":"Journal of the American College of Radiology","topic":"Artificial Intelligence in Healthcare and Education","field":"Medicine","cited_by":99,"is_retracted":false,"has_abstract":false,"ca_institutions":"York University","funders":"National Institute on Minority Health and Health Disparities; National Institute of Biomedical Imaging and Bioengineering; National Institute of Mental Health; Radiological Society of North America; National Bureau of Economic Research; National Heart, Lung, and Blood Institute; U.S. Department of Defense; National Cancer Institute; National Institutes of Health; National Science Foundation","keywords":"Artificial intelligence; Inference; Spurious relationship; Computer science; Machine learning; Metric (unit); Race (biology)","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.01873154,0.000913972,0.003644852,0.004441781,0.0005283213,0.003243835,0.002846425,0.003296074,0.004272484],"category_scores_gemma":[0.04590011,0.0006176108,0.002633158,0.00365668,0.003481674,0.003805844,0.00217619,0.003278658,0.0007777205],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002416871,"about_ca_system_score_gemma":0.006498821,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003716124,"about_ca_topic_score_gemma":0.008019926,"domain_scores_codex":[0.9908299,0.004235859,0.001380711,0.0007646946,0.00253589,0.0002529748],"domain_scores_gemma":[0.9442092,0.04611532,0.004194178,0.00103225,0.004066408,0.0003827684],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"systematic_review","study_design_scores_codex":[0.0002025539,0.00005356187,0.001980909,0.05226695,0.001611775,0.0001314039,0.0001938986,0.0003276285,0.0002030322,0.01275105,0.01080356,0.9194736],"study_design_scores_gemma":[0.0005072583,0.0007981329,0.02129986,0.227378,0.01184064,0.004436133,0.0008828368,0.001049232,0.002823167,0.08722359,0.6414956,0.0002655096],"study_design_candidate":"systematic_review","study_design_consensus":null,"genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.0002418486,0.9974534,0.00031996,0.001364777,0.0001336827,0.00001807291,0.00002584907,0.000006001786,0.0004363909],"genre_scores_gemma":[0.005724341,0.9906192,0.00109749,0.001801153,0.0004657453,0.00004416258,0.00004266438,0.000006668125,0.0001985107],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.9812685,"threshold_uncertainty_score":0.09906298,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.3374095107685531,"score_gpt":0.497149325622682,"score_spread":0.1597398148541289,"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."}}