{"id":"W7108079106","doi":"10.1007/s43681-025-00906-2","title":"The devil is in the details: using machine-learning to scrutinize “state-of-the-art” language models’ responses to public inquiries across 3 continents","year":2025,"lang":"en","type":"article","venue":"AI and Ethics","topic":"Artificial Intelligence in Healthcare and Education","field":"Medicine","cited_by":1,"is_retracted":false,"has_abstract":false,"ca_institutions":"École Nationale d'Administration Publique","funders":"Social Sciences and Humanities Research Council","keywords":"Indigenous; Sample (material); Field (mathematics); Quality (philosophy); Advice (programming); Comparative research; Variety (cybernetics); Public policy","routes":{"ca_aff":true,"ca_fund":true,"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":[0.002412302,0.00008210339,0.0001288499,0.00005857016,0.0005926165,0.00009998943,0.0001577319,0.00009206632,0.000003403617],"category_scores_gemma":[0.004751255,0.00004564226,0.00003326253,0.0005015003,0.0002272991,0.00007510273,0.00009919059,0.0006840405,0.000005209],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003238382,"about_ca_system_score_gemma":0.0003552284,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001864091,"about_ca_topic_score_gemma":0.004772943,"domain_scores_codex":[0.9987555,0.0003322275,0.0003029502,0.0001448673,0.0002162605,0.0002482264],"domain_scores_gemma":[0.9980175,0.001375676,0.00005470068,0.0002486628,0.0002476875,0.00005577829],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"qualitative","study_design_scores_codex":[0.0005905952,0.00006937949,0.2419328,0.0002330805,0.00003556321,0.000004401033,0.6292615,0.0002522519,0.0008486719,0.002997341,0.001907367,0.1218671],"study_design_scores_gemma":[0.0006019559,0.0007914273,0.05879923,0.003517884,0.0001767623,0.00005033676,0.5462592,0.05595617,0.0405952,0.0312739,0.2612361,0.0007419187],"study_design_candidate":"qualitative","study_design_consensus":"qualitative","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8360159,0.0007742991,0.00145449,0.1611267,0.0002135446,0.0002960171,0.000004227318,0.000009319819,0.0001054322],"genre_scores_gemma":[0.9752396,0.0001767933,0.0001278454,0.02326608,0.000040707,0.00001733484,0.000002100407,0.0000069485,0.001122604],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2593287,"threshold_uncertainty_score":0.568804,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2726920714253809,"score_gpt":0.5012050833285081,"score_spread":0.2285130119031272,"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."}}