{"id":"W4389763594","doi":"10.1080/03007995.2023.2295411","title":"The use of large language models in medicine: proceeding with caution","year":2023,"lang":"en","type":"article","venue":"Current Medical Research and Opinion","topic":"Artificial Intelligence in Healthcare and Education","field":"Medicine","cited_by":11,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia; University of British Columbia Hospital; Carleton University; University of Toronto","funders":"","keywords":"Medicine; Transparency (behavior); Consistency (knowledge bases); Liability; Internet privacy; Computer security; Artificial intelligence; Law","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":[0.1595244,0.001617488,0.002832208,0.003193762,0.003288076,0.01060328,0.008469796,0.009404132,0.002970682],"category_scores_gemma":[0.3762024,0.001746331,0.002325885,0.001590421,0.01663565,0.0157574,0.008567633,0.03364412,0.004860083],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002902466,"about_ca_system_score_gemma":0.006842145,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.008013935,"about_ca_topic_score_gemma":0.01382685,"domain_scores_codex":[0.8885484,0.0798882,0.008138826,0.00464054,0.01815321,0.0006308182],"domain_scores_gemma":[0.5872764,0.3353869,0.01016093,0.03551924,0.02878529,0.002871125],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.001065364,0.0001428872,0.009049654,0.005114642,0.001880209,0.00319358,0.01168549,0.005793989,0.002752662,0.3706021,0.3101328,0.2785866],"study_design_scores_gemma":[0.0001924349,0.0001534006,0.001919401,0.007572133,0.0004645014,0.002740867,0.002226238,0.01185469,0.003066971,0.6008216,0.3686122,0.000375396],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"commentary","genre_gemma":"commentary","genre_scores_codex":[0.004390162,0.03516027,0.2095013,0.7256429,0.01095187,0.000386855,0.0006421455,0.001273989,0.01205054],"genre_scores_gemma":[0.1324738,0.02586505,0.4147117,0.3890291,0.022898,0.001983357,0.0004176333,0.001209553,0.01141178],"genre_candidate":"commentary","genre_consensus":"commentary","teacher_disagreement_score":0.8404756,"threshold_uncertainty_score":0.8436556,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.6244928489176927,"score_gpt":0.5985246997722155,"score_spread":0.02596814914547729,"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."}}