{"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":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0005352963,0.0002921665,0.0003533991,0.0003406274,0.00009169507,0.00005630952,0.0004240972,0.0001376568,0.000003524589],"category_scores_gemma":[0.00002485853,0.0002522143,0.00003439344,0.000244728,0.00006289344,0.000100891,0.000290517,0.0003152461,0.00000156071],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002246315,"about_ca_system_score_gemma":0.0001798638,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00004490085,"about_ca_topic_score_gemma":0.0001842229,"domain_scores_codex":[0.9985994,0.000007190516,0.0003822375,0.0003908942,0.0002757586,0.0003445518],"domain_scores_gemma":[0.9991229,0.0002392187,0.00006184363,0.0004470061,0.00006339582,0.00006567912],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[6.711884e-7,0.000001565667,0.00003327749,0.0001685082,0.00001305338,0.00005897342,0.002183255,0.9347908,0.00007365693,0.0003083097,0.000003437724,0.06236451],"study_design_scores_gemma":[0.0001109118,0.0000113325,0.0001218915,0.0009996768,0.00000983792,0.0002890536,0.000001929428,0.9956791,0.001843987,0.0001621603,0.0004454774,0.0003246683],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01141349,0.00108558,0.9854603,0.00001590014,0.0007418041,0.0003300098,0.00002194233,0.0000908561,0.0008400718],"genre_scores_gemma":[0.8714233,0.000009731176,0.128304,0.00008054655,0.00009088853,0.0000070395,0.000001794142,0.00003057902,0.0000520968],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.8600098,"threshold_uncertainty_score":0.999993,"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."}}