{"id":"W4283170666","doi":"10.1145/3531146.3533088","title":"Taxonomy of Risks posed by Language Models","year":2022,"lang":"en","type":"article","venue":"2022 ACM Conference on Fairness, Accountability, and Transparency","topic":"Hate Speech and Cyberbullying Detection","field":"Computer Science","cited_by":604,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto","funders":"","keywords":"Misinformation; Taxonomy (biology); Harm; Risk analysis (engineering); Futures studies; Computer science; Risk management; Data science; Management science; Knowledge management; Engineering ethics; Psychology; Business; Artificial intelligence; Computer security; Social psychology; Engineering","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":[0.03159864,0.001876302,0.0008698409,0.008971202,0.007098341,0.01258973,0.003273756,0.007994696,0.003166432],"category_scores_gemma":[0.08222337,0.001032275,0.002142634,0.003522977,0.01633283,0.02388108,0.01323202,0.007793542,0.000791945],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.006929202,"about_ca_system_score_gemma":0.00667424,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004847505,"about_ca_topic_score_gemma":0.002538707,"domain_scores_codex":[0.9241995,0.03184341,0.00656043,0.003381646,0.03020209,0.003812977],"domain_scores_gemma":[0.8756186,0.07425313,0.01397791,0.01693241,0.01652305,0.002694993],"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":[0.00007163408,0.00009699293,0.01173439,0.0003847976,0.00006573895,0.0009807671,0.01619334,0.009468224,0.0007324404,0.9089332,0.002916417,0.04842203],"study_design_scores_gemma":[0.0000208423,0.0001081377,0.002261511,0.001106589,0.00008880118,0.002468324,0.01128418,0.02517925,0.001473445,0.8954215,0.06045306,0.0001343085],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.1465897,0.0081715,0.6380854,0.05345459,0.0005182718,0.001000893,0.0005506115,0.00126734,0.1503617],"genre_scores_gemma":[0.8288844,0.005505418,0.152171,0.002587718,0.000400749,0.000885728,0.0005562619,0.000238929,0.008769788],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.03159864,"threshold_uncertainty_score":0.1671116,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07387969715411008,"score_gpt":0.2846232755073295,"score_spread":0.2107435783532194,"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."}}