{"id":"W4411162779","doi":"10.1093/eurpub/ckaf072","title":"Evaluation of a large language model (ChatGPT) versus human researchers in assessing risk-of-bias and community engagement levels: a systematic review use-case analysis","year":2025,"lang":"en","type":"review","venue":"European Journal of Public Health","topic":"Artificial Intelligence in Healthcare and Education","field":"Medicine","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"","keywords":"Psychology; Risk analysis (engineering); Environmental health; 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":["metaresearch"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.6322416,0.003245851,0.008181987,0.01326045,0.002461573,0.008339355,0.004911606,0.004560041,0.003961608],"category_scores_gemma":[0.7506333,0.00312795,0.02331635,0.008616758,0.004479592,0.009100212,0.008420137,0.002995971,0.000599449],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.01390934,"about_ca_system_score_gemma":0.02350459,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00383578,"about_ca_topic_score_gemma":0.01065341,"domain_scores_codex":[0.2520649,0.5934432,0.09164859,0.01433936,0.04694826,0.001555741],"domain_scores_gemma":[0.1509869,0.7512379,0.04030067,0.02802705,0.02814247,0.00130494],"domain_codex":"methods","domain_gemma":"evaluation","domain_candidate":"evaluation","domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"systematic_review","study_design_scores_codex":[0.01528336,0.0007223143,0.06092069,0.3606487,0.1230818,0.0009663657,0.02211804,0.01044422,0.003740641,0.02109236,0.008741625,0.37224],"study_design_scores_gemma":[0.02468371,0.02099206,0.03982587,0.3118877,0.3422018,0.002530388,0.009232061,0.1082958,0.01444535,0.06675241,0.05734586,0.001807165],"study_design_candidate":"systematic_review","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"review","genre_scores_codex":[0.1077104,0.07108564,0.5583279,0.009153545,0.001408677,0.2377928,0.004716676,0.00211583,0.007688609],"genre_scores_gemma":[0.2964393,0.005345927,0.4542294,0.001942074,0.0001561859,0.2397905,0.001213888,0.0002812088,0.0006014945],"genre_candidate":"review","genre_consensus":null,"teacher_disagreement_score":0.3677584,"threshold_uncertainty_score":0.4535116,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.9190129423053119,"score_gpt":0.6486701219153373,"score_spread":0.2703428203899746,"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."}}