{"id":"W4200303997","doi":"10.1126/science.abi7176","title":"Filling gaps in trustworthy development of AI","year":2021,"lang":"en","type":"article","venue":"Science","topic":"Ethics and Social Impacts of AI","field":"Social Sciences","cited_by":50,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto; Mila - Quebec Artificial Intelligence Institute; McGill University","funders":"Engineering and Physical Sciences Research Council","keywords":"Trustworthiness; Audit; Computer science; Business; Computer security; Accounting","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":["sts"],"consensus_categories":[],"category_scores_codex":[0.1159716,0.0008605247,0.001239433,0.005530524,0.008614139,0.01645281,0.004020137,0.005917175,0.01044573],"category_scores_gemma":[0.2385199,0.001517717,0.0007915278,0.003205904,0.02204772,0.03977915,0.02087449,0.006964,0.002462155],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.009199994,"about_ca_system_score_gemma":0.04001443,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006947667,"about_ca_topic_score_gemma":0.005749652,"domain_scores_codex":[0.8910977,0.07276612,0.009184858,0.006030158,0.01628873,0.004632402],"domain_scores_gemma":[0.5852647,0.2391432,0.02966515,0.0773066,0.05785523,0.010765],"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":[0.0002237713,0.0003084064,0.01327845,0.001686122,0.00008962337,0.0006922117,0.03775757,0.006405015,0.001187876,0.7278135,0.008803264,0.2017541],"study_design_scores_gemma":[0.0000424077,0.0002088447,0.004675648,0.0038941,0.00004226328,0.0003385536,0.02584052,0.008058804,0.002273501,0.8570139,0.09751944,0.00009204846],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1305902,0.01496142,0.4037869,0.2440847,0.001634936,0.0007613527,0.0003869126,0.0009134898,0.2028801],"genre_scores_gemma":[0.9126574,0.004001343,0.07489724,0.002511426,0.0002604289,0.0003513653,0.0002039026,0.0002344203,0.004882362],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9913859,"threshold_uncertainty_score":0.613324,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04915249500649999,"score_gpt":0.400799252304235,"score_spread":0.351646757297735,"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."}}