{"id":"W4406827449","doi":"10.1016/j.jcjo.2025.01.001","title":"Assessing the performance of Microsoft Copilot, GPT-4 and Google Gemini in ophthalmology","year":2025,"lang":"en","type":"article","venue":"Canadian Journal of Ophthalmology","topic":"Retinal Imaging and Analysis","field":"Medicine","cited_by":20,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Toronto; Centre Hospitalier de l’Université de Montréal; Université de Montréal","funders":"Université de Montréal; University of Toronto","keywords":"Computer science; GCM transcription factors; Operations research; Aeronautics; Mathematics; Engineering; Biology","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":true,"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":[0.0005940244,0.0001075121,0.0004706955,0.0005452782,0.00008151269,0.00002432161,0.0001723107,0.00009335882,0.0001971664],"category_scores_gemma":[0.0002971127,0.00007835323,0.00008299284,0.0003926538,0.0004352105,0.00008846525,0.00001951074,0.0003824152,0.000001541827],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00008132158,"about_ca_system_score_gemma":0.001003661,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002081894,"about_ca_topic_score_gemma":0.00002849042,"domain_scores_codex":[0.9989351,0.000114291,0.0004824949,0.000126371,0.00007278323,0.0002689844],"domain_scores_gemma":[0.9990494,0.0001954455,0.0002258168,0.0001713289,0.0001827557,0.0001752682],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00003612066,0.0000228388,0.9885622,0.0000894366,0.00008333747,0.007770014,0.0001568203,0.00003250132,0.002307325,0.00004819895,0.0002169753,0.000674217],"study_design_scores_gemma":[0.0006955026,0.0004020956,0.7765228,0.0003842609,0.0001376136,0.2191266,0.0004753594,0.0003229261,0.00132272,0.0001436009,0.000391076,0.00007538768],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9821848,0.002767288,0.000005578651,0.002415418,0.0002020134,0.00006000703,0.000002271822,8.761358e-7,0.0123618],"genre_scores_gemma":[0.9982581,0.00001941271,0.0003726512,0.000118643,0.00004283134,0.000001220569,0.000001701689,0.000007743593,0.001177687],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2120394,"threshold_uncertainty_score":0.3195152,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02403652850803762,"score_gpt":0.3370729267759948,"score_spread":0.3130363982679571,"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."}}