{"id":"W4224940987","doi":"10.1145/3491102.3501999","title":"Human-AI Collaboration via Conditional Delegation: A Case Study of Content Moderation","year":2022,"lang":"en","type":"article","venue":"CHI Conference on Human Factors in Computing Systems","topic":"Explainable Artificial Intelligence (XAI)","field":"Computer Science","cited_by":128,"is_retracted":false,"has_abstract":true,"ca_institutions":"Microsoft (Canada)","funders":"National Science Foundation","keywords":"Computer science; Delegation; Testbed; Scalability; Moderation; Benchmark (surveying); Data science; Knowledge management; Artificial intelligence; Machine learning; World Wide Web","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":[0.0246697,0.001199987,0.001017328,0.001498316,0.00322748,0.002575435,0.003362521,0.003758485,0.003421553],"category_scores_gemma":[0.09664084,0.0004630717,0.0009715622,0.001721132,0.002633702,0.005660182,0.004892073,0.0035849,0.001031513],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001732025,"about_ca_system_score_gemma":0.001640758,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003649983,"about_ca_topic_score_gemma":0.005241745,"domain_scores_codex":[0.962741,0.03078279,0.001004611,0.002736943,0.002009129,0.0007255684],"domain_scores_gemma":[0.8065096,0.1572536,0.004807129,0.0242249,0.004343126,0.002861715],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"qualitative","study_design_scores_codex":[0.008187858,0.00810123,0.2448273,0.003326595,0.001040939,0.009324037,0.06005167,0.1896018,0.01951716,0.06166361,0.06702303,0.3273348],"study_design_scores_gemma":[0.001295299,0.002201333,0.02842992,0.000399293,0.0002604403,0.002414341,0.0127628,0.776148,0.01906379,0.06757714,0.08920889,0.0002387359],"study_design_candidate":"qualitative","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8520395,0.0009730789,0.1216508,0.004762838,0.0001881863,0.001215303,0.001830994,0.003290426,0.01404889],"genre_scores_gemma":[0.9234639,0.0001374756,0.07156998,0.000580498,0.00007497332,0.0004870907,0.001692967,0.000199522,0.00179349],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.0246697,"threshold_uncertainty_score":0.1304674,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1900154621451945,"score_gpt":0.369592466663248,"score_spread":0.1795770045180535,"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."}}