{"id":"W4229442586","doi":"10.1145/3531146.3533179","title":"The Road to Explainability is Paved with Bias: Measuring the Fairness of Explanations","year":2022,"lang":"en","type":"article","venue":"2022 ACM Conference on Fairness, Accountability, and Transparency","topic":"Explainable Artificial Intelligence (XAI)","field":"Computer Science","cited_by":65,"is_retracted":false,"has_abstract":true,"ca_institutions":"Vector Institute; University of Toronto","funders":"Government of Canada; Canadian Institute for Advanced Research; Vector Institute; Microsoft Research","keywords":"Computer science; Fidelity; Audit; Quality (philosophy); Machine learning; Artificial intelligence; High fidelity; Data science","routes":{"ca_aff":true,"ca_fund":true,"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":[0.06659783,0.0008683412,0.001461779,0.002896142,0.002674405,0.005170011,0.00176884,0.003633832,0.002743069],"category_scores_gemma":[0.4219501,0.0006111276,0.001301867,0.002089927,0.007614642,0.01120391,0.006755861,0.005094803,0.0003506087],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002973036,"about_ca_system_score_gemma":0.002857045,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004694396,"about_ca_topic_score_gemma":0.00253687,"domain_scores_codex":[0.9383346,0.04336707,0.003413344,0.005558355,0.00781709,0.001509559],"domain_scores_gemma":[0.4379481,0.4621044,0.03555987,0.04979374,0.01141992,0.003174125],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.003080867,0.0005194025,0.404851,0.001238563,0.001743894,0.0004952762,0.01439601,0.122083,0.003915726,0.2153347,0.007292385,0.2250492],"study_design_scores_gemma":[0.0002000585,0.0005256251,0.07345247,0.0006410375,0.0003875569,0.000444818,0.002821342,0.2841185,0.007644176,0.6210931,0.008410051,0.000261347],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5573982,0.002973649,0.4203902,0.009493044,0.0002219486,0.0002665733,0.001069156,0.0008224061,0.007364848],"genre_scores_gemma":[0.977473,0.0001358769,0.02134538,0.0003179923,0.0000746228,0.00006751696,0.0002575099,0.00009563297,0.0002325135],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.06659783,"threshold_uncertainty_score":0.3522072,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0904029058685791,"score_gpt":0.2902133418712273,"score_spread":0.1998104360026482,"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."}}