{"id":"W4407762808","doi":"10.1101/2025.02.17.25322388","title":"Collaborative intelligence in AI: Evaluating the performance of a council of AIs on the USMLE","year":2025,"lang":"en","type":"preprint","venue":"medRxiv","topic":"Big Data and Business Intelligence","field":"Business, Management and Accounting","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Health Care Foundation","funders":"","keywords":"Psychology; Medical education; Computer science; Medicine","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":[],"consensus_categories":[],"category_scores_codex":[0.09655967,0.00147877,0.001700697,0.002726742,0.002800746,0.004307753,0.003816419,0.003872997,0.003226484],"category_scores_gemma":[0.2909123,0.0008115643,0.001209837,0.0017937,0.002170631,0.006138409,0.00887489,0.003965271,0.001281315],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003833485,"about_ca_system_score_gemma":0.00628487,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01518011,"about_ca_topic_score_gemma":0.009901199,"domain_scores_codex":[0.9309528,0.0513443,0.003659283,0.005837206,0.006450132,0.001756235],"domain_scores_gemma":[0.5969542,0.3261997,0.02089079,0.02340445,0.02167544,0.0108754],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"observational","study_design_scores_codex":[0.01532701,0.008260457,0.2311596,0.001111009,0.001665972,0.0004288205,0.01245049,0.3709652,0.003978979,0.007253336,0.009534745,0.3378644],"study_design_scores_gemma":[0.001122067,0.007826943,0.02627865,0.000201057,0.0004268181,0.0001266644,0.0038338,0.9414892,0.004715382,0.006642275,0.007131098,0.0002061113],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9614072,0.0003439137,0.02551628,0.001397043,0.0001440921,0.001087555,0.0002572368,0.0009863706,0.00886031],"genre_scores_gemma":[0.9591093,0.00007329785,0.03822111,0.00029779,0.0000491764,0.0007482083,0.0005015971,0.00007336205,0.0009262823],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.09655967,"threshold_uncertainty_score":0.5106624,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1489686306185597,"score_gpt":0.3471293997530806,"score_spread":0.1981607691345209,"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."}}