{"id":"W4415757988","doi":"10.70777/si.v2i5.16331","title":"A Grading Rubric for AI Safety Frameworks","year":2025,"lang":"","type":"article","venue":"SuperIntelligence - Robotics - Safety & Alignment","topic":"Ethics and Social Impacts of AI","field":"Social Sciences","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Institute on Governance","funders":"","keywords":"Rubric; Grading (engineering); Warrant; Delphi method; Delphi","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":[0.05584512,0.002018233,0.001494925,0.0241098,0.006567659,0.006827294,0.003982022,0.002742099,0.01173818],"category_scores_gemma":[0.1763208,0.0008742653,0.002362447,0.009630477,0.003900697,0.006393262,0.00710105,0.00398588,0.00794703],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.01286242,"about_ca_system_score_gemma":0.01814733,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.008166137,"about_ca_topic_score_gemma":0.009794668,"domain_scores_codex":[0.9337733,0.02563093,0.01497336,0.001783676,0.02115138,0.002687278],"domain_scores_gemma":[0.7889575,0.04059212,0.0121925,0.01128783,0.1413447,0.005625265],"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":[0.0002901971,0.0009246144,0.009070924,0.002932302,0.00005339001,0.0004172296,0.01858831,0.004478165,0.00727897,0.09633476,0.2460925,0.6135386],"study_design_scores_gemma":[0.0001751799,0.001060716,0.02815646,0.004105173,0.00007607156,0.001048958,0.0309089,0.01979569,0.007203299,0.07651063,0.8302516,0.0007073666],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.07481391,0.00301891,0.5753301,0.0162561,0.005825372,0.04991471,0.005983022,0.01052614,0.2583317],"genre_scores_gemma":[0.1382713,0.001955573,0.7889082,0.002343858,0.0006846609,0.03296959,0.007856151,0.001070795,0.02593995],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.05584512,"threshold_uncertainty_score":0.2953408,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0323729125467903,"score_gpt":0.3708621776750627,"score_spread":0.3384892651282724,"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."}}