{"id":"W4417101019","doi":"10.70777/si.v2i4.16671","title":"International AI Safety Report 2025: Second Key Update: Technical Safeguards and Risk Management","year":2025,"lang":"","type":"article","venue":"SuperIntelligence - Robotics - Safety & Alignment","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto; Mila - Quebec Artificial Intelligence Institute","funders":"","keywords":"Key (lock); Risk management; SAFER; Risk assessment; Safeguard; Corporate governance; Risk governance","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true,"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":[0.02602913,0.003130724,0.002287309,0.006809821,0.00322963,0.01408024,0.007126012,0.016672,0.0630223],"category_scores_gemma":[0.06618345,0.001245554,0.002364158,0.003268813,0.002824303,0.008087251,0.004919813,0.01226967,0.07579745],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.01307184,"about_ca_system_score_gemma":0.03927463,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.06550834,"about_ca_topic_score_gemma":0.0453088,"domain_scores_codex":[0.9621258,0.00398751,0.002814587,0.001291876,0.02618717,0.003593024],"domain_scores_gemma":[0.890845,0.009949615,0.004739937,0.003873447,0.08644357,0.004148432],"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":[0.00002937623,0.00002635917,0.0001268013,0.0002542287,0.00001070936,0.00003755479,0.00004099547,0.000233661,0.0001944638,0.003388875,0.9713866,0.02427045],"study_design_scores_gemma":[0.000009011786,0.0000256762,0.0002666982,0.0003778465,0.00001145774,0.00003520956,0.00004459347,0.0001246119,0.0002329023,0.001230648,0.9976181,0.0000230137],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"other","genre_gemma":"other","genre_scores_codex":[0.001451457,0.03263022,0.03177185,0.25316,0.172208,0.002487118,0.03141798,0.0102199,0.4646535],"genre_scores_gemma":[0.0286417,0.05172183,0.04873934,0.1235161,0.04837313,0.004863664,0.07945286,0.005034108,0.6096572],"genre_candidate":"other","genre_consensus":"other","teacher_disagreement_score":0.06550834,"threshold_uncertainty_score":0.2108306,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.009785592565935903,"score_gpt":0.2868525432628055,"score_spread":0.2770669506968696,"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."}}