{"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":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","sts","scholarly_communication","research_integrity","insufficient_payload"],"consensus_categories":["research_integrity"],"category_scores_codex":[0.005357329,0.001116984,0.001514713,0.0003855668,0.005202257,0.001344671,0.002132237,0.002564434,0.0009360673],"category_scores_gemma":[0.00280202,0.001254028,0.001243946,0.002148629,0.002060518,0.0009698975,0.0006891242,0.002822229,0.0001543385],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002537464,"about_ca_system_score_gemma":0.002403239,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001726334,"about_ca_topic_score_gemma":0.0008915103,"domain_scores_codex":[0.9904405,0.0007324198,0.002458148,0.001727639,0.00177436,0.002866951],"domain_scores_gemma":[0.9928394,0.003150532,0.0004562662,0.001248757,0.001314406,0.0009906531],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0003935279,0.0006536615,0.0007011672,0.0003861628,0.0007293353,0.00001497408,0.02958393,0.0426941,0.0001157383,0.8947176,0.002545314,0.02746446],"study_design_scores_gemma":[0.001173043,0.0009360185,0.0005132201,0.002605356,0.001037295,0.000003978033,0.0498033,0.02178906,0.001735515,0.5255078,0.3921654,0.002730047],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0005486151,0.003590551,0.820213,0.1267953,0.01003238,0.003795617,0.0002591919,0.0002190277,0.03454631],"genre_scores_gemma":[0.9229221,0.03941546,0.01327055,0.0126065,0.001200078,0.0001405099,0.00006550243,0.0001412426,0.01023806],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9223735,"threshold_uncertainty_score":0.9999772,"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."}}