{"id":"W4410932950","doi":"10.1186/s13750-025-00362-9","title":"Evaluating generative AI for qualitative data extraction in community-based fisheries management literature","year":2025,"lang":"en","type":"article","venue":"Environmental Evidence","topic":"Computational and Text Analysis Methods","field":"Social Sciences","cited_by":9,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Centre for Marine Socioecology; Canada Excellence Research Chairs, Government of Canada; Commonwealth Scientific and Industrial Research Organisation","keywords":"Context (archaeology); Computer science; Data science; Data extraction; Generative grammar; Systematic review; Knowledge management; Artificial intelligence; MEDLINE; Political science","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004533591,0.00008894479,0.0001208176,0.00009912468,0.0006609038,0.0001120382,0.0003720769,0.00004313145,0.00007246208],"category_scores_gemma":[0.0004773619,0.00009065446,0.00004319759,0.0003250501,0.0001819607,0.0006619443,0.0001475339,0.0001965843,0.000004398245],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002390888,"about_ca_system_score_gemma":0.00004937654,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0003413385,"about_ca_topic_score_gemma":0.0007635273,"domain_scores_codex":[0.9966435,0.002465153,0.0002121334,0.0002457161,0.0002962499,0.000137253],"domain_scores_gemma":[0.9970451,0.002569345,0.00007814363,0.0002657532,0.00001578559,0.00002586018],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"qualitative","study_design_scores_codex":[0.0006417927,0.001359195,0.01583922,0.0006199753,0.0003910568,0.00001677703,0.3485876,0.02166061,0.006727192,0.03835489,0.006671559,0.5591301],"study_design_scores_gemma":[0.002969951,0.0005132621,0.1275468,0.003791256,0.0005183431,7.803612e-7,0.3138935,0.2211917,0.001861687,0.2169179,0.1094726,0.001322328],"study_design_candidate":"qualitative","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3670461,0.00386969,0.5958101,0.02643845,0.0005766118,0.002292595,0.000204823,0.00007931729,0.003682376],"genre_scores_gemma":[0.8889546,0.0001584678,0.1067704,0.00151472,0.00004968348,0.0001280865,0.00029047,0.000005024093,0.002128504],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.5578077,"threshold_uncertainty_score":0.5083205,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2878373345554333,"score_gpt":0.5558607988604336,"score_spread":0.2680234643050003,"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."}}