{"id":"W4409716856","doi":"10.1111/jels.12413","title":"Hallucination‐Free? Assessing the Reliability of Leading <scp>AI</scp> Legal Research Tools","year":2025,"lang":"en","type":"article","venue":"Journal of Empirical Legal Studies","topic":"Artificial Intelligence in Law","field":"Social Sciences","cited_by":92,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Stanford Law School; University of Toronto; Fordham University","keywords":"Hallucinating; Automatic summarization; Legal psychology; Typology; Lexis; Computer science; Legal case; Legal research; Psychology; Data science; Artificial intelligence; Social psychology; Law; Political science; Sociology; Linguistics","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":["metaresearch","sts"],"consensus_categories":[],"category_scores_codex":[0.0127364,0.0001419198,0.0005425516,0.0002389818,0.001812799,0.0004803458,0.001068165,0.0001399667,0.00001870298],"category_scores_gemma":[0.1031283,0.00009249943,0.0002486983,0.001489321,0.002706222,0.001471089,0.0003936469,0.001181092,0.00001092485],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006207001,"about_ca_system_score_gemma":0.0009959702,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0007257888,"about_ca_topic_score_gemma":0.0008106252,"domain_scores_codex":[0.9949276,0.001558599,0.001095937,0.0002290724,0.001665056,0.0005236879],"domain_scores_gemma":[0.9794127,0.01587167,0.0004549226,0.0003205662,0.003838593,0.0001015421],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00006992786,0.0007763413,0.09836544,0.0001759487,0.0006017438,0.00008555009,0.05089012,0.001692203,0.001047003,0.3712644,0.4528454,0.02218593],"study_design_scores_gemma":[0.0001756334,0.0002567167,0.01078339,0.0004089488,0.00009748155,0.000005351978,0.07641403,0.0001941497,0.004013919,0.07476898,0.8327911,0.0000902322],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.688195,0.004167591,0.00244533,0.1309343,0.002193033,0.0004326283,0.000004741816,0.00003488769,0.1715925],"genre_scores_gemma":[0.9935347,0.0003503265,0.0007446281,0.0006458233,0.0006524799,0.000006327693,1.221254e-7,0.000008167616,0.004057428],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3799458,"threshold_uncertainty_score":0.9994867,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.3252983431727126,"score_gpt":0.5757539876260013,"score_spread":0.2504556444532887,"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."}}