{"id":"W4387986344","doi":"10.2139/ssrn.4587092","title":"Lawyers Should Not Trust AI: A call for an Open-source Legal Language Model","year":2023,"lang":"en","type":"article","venue":"SSRN Electronic Journal","topic":"Artificial Intelligence in Law","field":"Social Sciences","cited_by":2,"is_retracted":false,"has_abstract":false,"ca_institutions":"Queen's University","funders":"","keywords":"Computer science; Law; Linguistics; Psychology; Political science; Business; Philosophy","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":"codex-gemma-dda1882f352a","candidate_categories":["sts"],"consensus_categories":[],"category_scores_codex":[0.00515671,0.0001696056,0.0002279137,0.0001295704,0.001428348,0.000562656,0.001359463,0.0001677995,0.0000583482],"category_scores_gemma":[0.0003806024,0.0001639254,0.0001332757,0.0004496381,0.0002384131,0.001046568,0.0001125212,0.001326554,0.0001101675],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009745377,"about_ca_system_score_gemma":0.003769884,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.007313993,"about_ca_topic_score_gemma":0.1133932,"domain_scores_codex":[0.9956212,0.0001907886,0.0003460161,0.0003263973,0.0005556688,0.002959909],"domain_scores_gemma":[0.9990938,0.0001386106,0.000132606,0.000230169,0.0001545539,0.000250268],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0001738517,0.00008115323,0.00007685141,0.000003343424,0.0000552724,0.00000671371,0.01455458,0.007465423,0.001311314,0.9173705,0.002719552,0.05618147],"study_design_scores_gemma":[0.0007654696,0.001164242,0.00001447529,0.00003498252,0.0001005273,0.00006204698,0.1392862,0.1346647,0.001513156,0.4440621,0.2774041,0.0009279667],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6975954,0.0004510159,0.2078705,0.03588248,0.001391505,0.002271291,0.00006153743,0.0009033784,0.05357288],"genre_scores_gemma":[0.9591717,0.000479856,0.0002552709,0.001293624,0.0008800646,0.00003903492,0.000009049458,0.00005577822,0.03781563],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4733084,"threshold_uncertainty_score":0.9998717,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.09281858513244648,"score_gpt":0.4159521215775243,"score_spread":0.3231335364450778,"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."}}