{"id":"W4413413852","doi":"10.1007/978-3-032-01241-8_14","title":"Balancing the Risks and Benefits of Using Large Language Models to Support Assurance Case Development","year":2025,"lang":"en","type":"book-chapter","venue":"Lecture notes in computer science","topic":"Safety Systems Engineering in Autonomy","field":"Engineering","cited_by":2,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Toronto; Critical Systems Labs; University of Victoria","funders":"","keywords":"Computer science; Software engineering; Risk analysis (engineering); Operations research; Engineering","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":[],"consensus_categories":[],"category_scores_codex":[0.02213932,0.0007769939,0.0005334327,0.002079285,0.0009171802,0.006297626,0.002718285,0.00195748,0.003681358],"category_scores_gemma":[0.1301558,0.001162425,0.0007838398,0.001248342,0.00181489,0.01090947,0.006146993,0.00355968,0.001645837],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001697605,"about_ca_system_score_gemma":0.002893342,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001888954,"about_ca_topic_score_gemma":0.003573221,"domain_scores_codex":[0.977944,0.01226339,0.001015534,0.0007972723,0.00730043,0.0006793249],"domain_scores_gemma":[0.773734,0.1916301,0.004766282,0.02239801,0.006105121,0.001366424],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.001046158,0.0007739507,0.01162833,0.0006433068,0.0001761649,0.000886086,0.00494623,0.09778705,0.02341943,0.1861418,0.008123025,0.6644284],"study_design_scores_gemma":[0.0002392602,0.0008388359,0.004749062,0.001166884,0.0003525868,0.001462407,0.002558089,0.6492295,0.02595496,0.2597266,0.05352346,0.0001981851],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"other","genre_scores_codex":[0.1029243,0.0008498019,0.8524291,0.005942372,0.0001433181,0.000653791,0.0002924058,0.004205085,0.03255973],"genre_scores_gemma":[0.5164481,0.0004862528,0.4765978,0.0005512299,0.00005743783,0.0004399397,0.0004445698,0.0008129611,0.004161773],"genre_candidate":"other","genre_consensus":null,"teacher_disagreement_score":0.02213932,"threshold_uncertainty_score":0.1170853,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02598692833681027,"score_gpt":0.2424217880257773,"score_spread":0.216434859688967,"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."}}