{"id":"W4416961865","doi":"10.1109/pst65910.2025.11268824","title":"G-STAR: A Threat Modeling Framework for General-Purpose AI Systems","year":2025,"lang":"en","type":"article","venue":"","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"National Research Council Canada","funders":"","keywords":"Key (lock); Software deployment; Threat model; Work (physics); Focus (optics); Taxonomy (biology)","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":[],"consensus_categories":[],"category_scores_codex":[0.0003820606,0.0001789723,0.0002491492,0.0001520422,0.0002771501,0.0003378341,0.0009924215,0.0001482139,0.000007671807],"category_scores_gemma":[0.0002407767,0.0001577479,0.0001040151,0.0004346167,0.00001838835,0.0004063322,0.0003955382,0.0002958005,0.00001478186],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00008144545,"about_ca_system_score_gemma":0.0001154117,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001747693,"about_ca_topic_score_gemma":0.000004936905,"domain_scores_codex":[0.998569,0.00006722299,0.0002899693,0.0005058012,0.0001993918,0.0003685758],"domain_scores_gemma":[0.9987437,0.0002842338,0.0000557455,0.0007075157,0.0001446515,0.00006418333],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000005003683,0.000008675835,0.00009704174,0.00002672529,0.00001507382,0.000001490721,0.0000755827,0.4084443,0.00001713183,0.5883319,0.0002841713,0.002692893],"study_design_scores_gemma":[0.0002252892,0.00001972424,0.000007495905,0.00007839862,0.00001065492,0.000002021498,0.00003489439,0.8769337,0.00004315846,0.1206087,0.001882531,0.0001533783],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.001564324,0.0003481332,0.9906006,0.00233481,0.001686071,0.0004640484,0.000001163169,0.0004684696,0.002532314],"genre_scores_gemma":[0.281923,0.000007668966,0.7134303,0.001262765,0.0002231926,0.00008825862,0.000002381124,0.00001401043,0.003048404],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.4684894,"threshold_uncertainty_score":0.643277,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02117561319398679,"score_gpt":0.3216072551423613,"score_spread":0.3004316419483745,"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."}}