{"id":"W4403887506","doi":"10.1007/s00146-024-02103-x","title":"Automating public policy: a comparative study of conversational artificial intelligence models and human expertise in crafting briefing notes","year":2024,"lang":"en","type":"article","venue":"AI & Society","topic":"Ethics and Social Impacts of AI","field":"Social Sciences","cited_by":6,"is_retracted":false,"has_abstract":false,"ca_institutions":"Dalhousie University; École Nationale d'Administration Publique","funders":"Social Sciences and Humanities Research Council of Canada","keywords":"Performing arts; Computer science; Knowledge management; Artificial intelligence; Engineering; Visual arts; Art","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00140176,0.00009809106,0.0002023924,0.00006212113,0.0007072111,0.0004227573,0.000126143,0.0001152133,0.00001594081],"category_scores_gemma":[0.0003898574,0.0001034885,0.0000736013,0.000506368,0.0004386427,0.0008101632,0.00007349055,0.0003193935,0.000001138661],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001904301,"about_ca_system_score_gemma":0.0005001551,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.03469615,"about_ca_topic_score_gemma":0.01242491,"domain_scores_codex":[0.9985622,0.000194756,0.0003367129,0.0002264474,0.0004060677,0.0002738626],"domain_scores_gemma":[0.9989876,0.0006001706,0.00007632269,0.00006722824,0.0001870389,0.00008171075],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"qualitative","study_design_scores_codex":[0.000001009304,0.00009941159,0.002194006,0.00001897164,0.00003467996,0.000001046327,0.7174565,0.0001399885,0.0002708792,0.276067,0.00004432452,0.003672183],"study_design_scores_gemma":[0.0001695412,0.0001473592,0.005211304,0.0001998951,0.00002110502,3.934649e-7,0.5505299,0.1808971,0.0001789788,0.2622927,0.00004228713,0.0003094886],"study_design_candidate":"qualitative","study_design_consensus":"qualitative","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9830005,0.0001991134,0.001813737,0.01371859,0.0000602478,0.0002513969,0.000004000918,0.00007434591,0.0008781092],"genre_scores_gemma":[0.9990129,0.00003263482,0.0003031797,0.0004400111,0.0001773803,0.00001206712,0.000001957753,0.000008170015,0.0000117344],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1807571,"threshold_uncertainty_score":0.9717319,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2424215892621185,"score_gpt":0.4660790036172732,"score_spread":0.2236574143551547,"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."}}