{"meta":{"page":1,"per_page":50,"max_per_page":100,"total":14,"total_is_capped":false,"direct_labels_cover":0,"predictions_cover":14,"direct_label_status":"direct model label, unvalidated","prediction_status":"machine_predicted_unvalidated (Codex and Gemma teacher distillation)","score_status":"score_only:v0-immature-baseline (scores rank; they never assert a category)","snapshot":{"source":"OpenAlex, pinned release, all 482 partitions","release":"2026-06-24","frame_built":"2026-07-12"},"query_hash":"fed833773f56","filters":{"venue":"SuperIntelligence - Robotics - Safety & Alignment"}},"results":[{"id":"W4411012657","doi":"10.70777/si.v2i2.14755","title":"The First International AI Safety Report","year":2025,"lang":"en","type":"article","venue":"SuperIntelligence - Robotics - Safety & Alignment","topic":"Ethics and Social Impacts of AI","field":"Social Sciences","cited_by":27,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal","funders":"Mozilla Foundation; Royal Society; DeepMind","keywords":"Forensic engineering; Engineering","retraction":null,"screen_n_in":null,"score":{"opus":0.02293576748403459,"gpt":0.361102392748586,"spread":0.3381666252645514,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["sts"],"consensus_categories":[],"category_scores_codex":[0.002975593,0.0002307411,0.0002483101,0.00006971537,0.003337622,0.0005622286,0.001171456,0.0002358385,0.0003117075],"category_scores_gemma":[0.001470247,0.0001943494,0.0002244197,0.0004054525,0.0009558807,0.0003383554,0.0003171214,0.0004655336,0.00009485552],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008598411,"about_ca_system_score_gemma":0.0007603238,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002228525,"about_ca_topic_score_gemma":0.006232953,"domain_scores_codex":[0.996843,0.000158806,0.00079642,0.0004397736,0.001076006,0.000686031],"domain_scores_gemma":[0.9978321,0.0007078899,0.000141704,0.0005664433,0.0005487502,0.0002031679],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0000716083,0.00011714,0.004081993,0.00001108729,0.0001901844,0.00005158601,0.007026908,0.007174189,0.0000137854,0.966854,0.004090483,0.01031703],"study_design_scores_gemma":[0.0001037077,0.00002837669,0.001417399,0.00007286249,0.0000289141,0.000003501274,0.007287334,0.0003521934,0.0001924271,0.04420989,0.9460683,0.0002351233],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"commentary","genre_gemma":"empirical","genre_scores_codex":[0.0005412044,0.0005409989,0.07639238,0.5307176,0.008911761,0.0009013967,0.00002936283,0.0001943018,0.381771],"genre_scores_gemma":[0.9360189,0.01624316,0.001477622,0.007497282,0.0008339079,0.00004147153,0.00003436574,0.00003597739,0.03781733],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9419778,"threshold_uncertainty_score":0.9979599,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4412505633","doi":"10.70777/si.v2i3.15063","title":"Darwin Gödel Machine: Open-Ended Evolution of Self-Improving Agents","year":2025,"lang":"en","type":"article","venue":"SuperIntelligence - Robotics - Safety & Alignment","topic":"Evolutionary Algorithms and Applications","field":"Computer Science","cited_by":12,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Vector Institute; Canadian Institute for Advanced Research; University of British Columbia","funders":"Natural Sciences and Engineering Research Council of Canada; Government of Canada; Canadian Institute for Advanced Research","keywords":"Darwin (ADL); Computer science; Software engineering","retraction":null,"screen_n_in":null,"score":{"opus":0.01750147175455305,"gpt":0.2803470702291549,"spread":0.2628455984746019,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0005214629,0.0002854732,0.0003376216,0.0001691227,0.0004089315,0.0001400217,0.002503368,0.0001069363,0.00007027271],"category_scores_gemma":[0.00003363841,0.0002927787,0.0001366868,0.000940708,0.0001007462,0.0005766518,0.001641247,0.000197233,0.00005809612],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000611991,"about_ca_system_score_gemma":0.0004134976,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0003715551,"about_ca_topic_score_gemma":0.00001814276,"domain_scores_codex":[0.9973931,0.00009314172,0.0008180709,0.0007510316,0.0004798851,0.0004647527],"domain_scores_gemma":[0.9981717,0.00009567905,0.0001755791,0.001227956,0.0001866233,0.000142402],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0000175139,0.0008327861,0.0009451335,0.00008521036,0.000114277,0.000005175771,0.0005003161,0.09974301,0.002580495,0.8720186,0.0005345713,0.02262285],"study_design_scores_gemma":[0.000488893,0.000162206,0.007241849,0.0001277172,0.00006323025,0.00001462099,0.0003885903,0.946618,0.008902059,0.02680484,0.008665624,0.0005223814],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0006381384,0.0004769808,0.989634,0.001785999,0.0005485562,0.0008136421,0.00003089257,0.0001551281,0.005916605],"genre_scores_gemma":[0.6902934,0.0002592152,0.3083371,0.0001842254,0.00003418173,0.00007816588,0.00002054999,0.00001419162,0.0007790382],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.846875,"threshold_uncertainty_score":0.9999524,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4413417613","doi":"10.70777/si.v2i5.15569","title":"Superintelligent Agents Pose Catastrophic Risks: Can Scientist AI Offer a Safer Path?","year":2025,"lang":"en","type":"article","venue":"SuperIntelligence - Robotics - Safety & Alignment","topic":"Space Science and Extraterrestrial Life","field":"Physics and Astronomy","cited_by":11,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"McGill University; Mila - Quebec Artificial Intelligence Institute","funders":"","keywords":"SAFER; Path (computing); Computer science; Artificial intelligence; Risk analysis (engineering); Computer security; Medicine","retraction":null,"screen_n_in":null,"score":{"opus":0.02974086538017891,"gpt":0.3159395806666005,"spread":0.2861987152864216,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0006294249,0.0006749246,0.0006342855,0.0002823609,0.0007349091,0.000423936,0.001164023,0.0001296149,0.002451066],"category_scores_gemma":[0.00003068806,0.0006322396,0.0004060161,0.001017204,0.0005106073,0.00031999,0.0005097779,0.0004625345,0.000595393],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003780815,"about_ca_system_score_gemma":0.0005758604,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002638886,"about_ca_topic_score_gemma":0.0001050039,"domain_scores_codex":[0.9952983,0.0001588989,0.001073751,0.001221065,0.0008840425,0.001363899],"domain_scores_gemma":[0.9978737,0.00008779901,0.0001392066,0.001188024,0.0001954025,0.0005158925],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"not_applicable","study_design_scores_codex":[0.000756804,0.003396174,0.2272124,0.0002181705,0.001350044,0.0001633831,0.00677656,0.3982562,0.01180731,0.2602153,0.019957,0.06989064],"study_design_scores_gemma":[0.006101593,0.002168883,0.05550541,0.002585726,0.001663148,0.00005859929,0.04256151,0.07147599,0.2930809,0.06554618,0.4496382,0.009613909],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2351901,0.0006071493,0.7084872,0.01368698,0.01493147,0.003896209,0.001330609,0.0003707793,0.02149957],"genre_scores_gemma":[0.9927804,0.0001369422,0.001537198,0.0008080727,0.0004376436,0.00008408337,0.0002014807,0.00004725656,0.003966924],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.7575903,"threshold_uncertainty_score":0.9996129,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4413130428","doi":"10.70777/si.v2i5.15503","title":"The Singapore Consensus on Global AI Safety Research Priorities","year":2025,"lang":"en","type":"article","venue":"SuperIntelligence - Robotics - Safety & Alignment","topic":"Artificial Intelligence in Healthcare and Education","field":"Medicine","cited_by":3,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal; Mila - Quebec Artificial Intelligence Institute","funders":"","keywords":"Pace; Autonomy; Trustworthiness; Software deployment; Public relations; Computer science; Political science; Business; Computer security; Law","retraction":null,"screen_n_in":null,"score":{"opus":0.1626971727495097,"gpt":0.4808835490017747,"spread":0.318186376252265,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","sts"],"consensus_categories":[],"category_scores_codex":[0.002295268,0.0003434185,0.0004277884,0.0001527796,0.001753958,0.0001559232,0.0004458436,0.0002539395,0.0001497667],"category_scores_gemma":[0.001140117,0.0002653014,0.0001980771,0.0009883613,0.001012295,0.00005876857,0.0001883632,0.0008524894,0.0003920115],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001964097,"about_ca_system_score_gemma":0.001675566,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0009594879,"about_ca_topic_score_gemma":0.0003679093,"domain_scores_codex":[0.9954829,0.000375802,0.001213886,0.0006916339,0.001124665,0.001111091],"domain_scores_gemma":[0.9958703,0.001703557,0.00008603079,0.001128816,0.0008869183,0.0003244025],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.003567086,0.001015294,0.02026201,0.0003235944,0.0003165886,0.00009573276,0.003013611,0.01710225,0.0002985689,0.5835154,0.01474124,0.3557486],"study_design_scores_gemma":[0.0004670183,0.004327139,0.0139913,0.003268882,0.0003185006,0.0001952063,0.07469968,0.01155197,0.06379456,0.1279959,0.6980184,0.001371436],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"commentary","genre_gemma":"empirical","genre_scores_codex":[0.1685465,0.00948577,0.05252273,0.6261094,0.0169887,0.009548399,0.0002049767,0.0007097792,0.1158837],"genre_scores_gemma":[0.9860605,0.003938597,0.001251785,0.002770405,0.0004372948,0.00007265815,0.00003461523,0.00003081055,0.005403306],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.817514,"threshold_uncertainty_score":0.9999799,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4413207531","doi":"10.70777/si.v2i5.15509","title":"In Which Areas of Technical AI Safety Could Geopolitical Rivals Cooperate?","year":2025,"lang":"en","type":"article","venue":"SuperIntelligence - Robotics - Safety & Alignment","topic":"Ethics and Social Impacts of AI","field":"Social Sciences","cited_by":2,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Mila - Quebec Artificial Intelligence Institute; Institute on Governance","funders":"","keywords":"Harm; Geopolitics; Risk analysis (engineering); China; Business; Computer security; Political science; Computer science","retraction":null,"screen_n_in":null,"score":{"opus":0.02611724820704924,"gpt":0.3797726841182025,"spread":0.3536554359111533,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.002899456,0.0003196,0.0006906248,0.0001911483,0.0005723339,0.0001324962,0.0007181909,0.0006033666,0.0006751489],"category_scores_gemma":[0.00212717,0.0003282781,0.000188318,0.001243737,0.001261694,0.0003788075,0.0002971747,0.000805946,0.00004736063],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009654192,"about_ca_system_score_gemma":0.001605713,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005155359,"about_ca_topic_score_gemma":0.007383995,"domain_scores_codex":[0.9956986,0.0004768972,0.00119873,0.0005539489,0.001031812,0.001039955],"domain_scores_gemma":[0.9973533,0.0008274058,0.000102836,0.0005430911,0.0008050187,0.0003684073],"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":[0.0001057317,0.0004327633,0.003194153,0.00005523693,0.00005483849,0.00001819789,0.005023062,0.00750428,0.0003587435,0.9822301,0.0004157167,0.0006072025],"study_design_scores_gemma":[0.002568809,0.001143249,0.0287906,0.002378716,0.000426836,0.00001792901,0.05147797,0.003820886,0.0212891,0.5863668,0.2982924,0.003426723],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"other","genre_gemma":"empirical","genre_scores_codex":[0.04661713,0.001395341,0.1082755,0.3430499,0.005003192,0.004628888,0.0002959059,0.0004406001,0.4902936],"genre_scores_gemma":[0.9906809,0.00291737,0.001299114,0.002614557,0.000140807,0.0000200067,0.00001373028,0.00002420716,0.002289331],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9440637,"threshold_uncertainty_score":0.9999169,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4410907699","doi":"10.70777/si.v2i2.14611","title":"Pitfalls of Evidence-Based AI Policy","year":2025,"lang":"en","type":"article","venue":"SuperIntelligence - Robotics - Safety & Alignment","topic":"Artificial Intelligence in Healthcare and Education","field":"Medicine","cited_by":2,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"","keywords":"Computer science; Political science","retraction":null,"screen_n_in":null,"score":{"opus":0.2312626749960818,"gpt":0.4610154853350626,"spread":0.2297528103389808,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.000685824,0.0003036887,0.0005516469,0.0004509739,0.0001737948,0.00003012771,0.0003239842,0.000209132,0.0004227723],"category_scores_gemma":[0.0008677357,0.0002575249,0.0002592903,0.001062489,0.0003126368,0.0001709242,0.00008980784,0.000311153,0.0001487973],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007354729,"about_ca_system_score_gemma":0.002306099,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001886081,"about_ca_topic_score_gemma":0.00005701034,"domain_scores_codex":[0.9969037,0.0001147517,0.001281415,0.0005217356,0.0005948393,0.0005835788],"domain_scores_gemma":[0.997588,0.0005362899,0.0001465629,0.0008845735,0.0005705809,0.0002739426],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.002769728,0.003071001,0.142987,0.003551389,0.000541561,0.00005484384,0.005191305,0.1881517,0.02723128,0.1914473,0.003367465,0.4316354],"study_design_scores_gemma":[0.0002299781,0.001687343,0.01039031,0.005697353,0.0004034737,0.00002429928,0.004393253,0.01909823,0.9196876,0.01598909,0.02163412,0.0007649619],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0740966,0.004155743,0.7119298,0.1970636,0.00330363,0.00346655,0.00003918934,0.0002758487,0.005669023],"genre_scores_gemma":[0.9847952,0.002359651,0.005429039,0.005394419,0.0002679825,0.00004926765,0.00002550332,0.00002786224,0.001651044],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9106987,"threshold_uncertainty_score":0.9999877,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4412505644","doi":"10.70777/si.v2i3.15155","title":"Trends in Frontier AI Model Count: A Forecast to 2028","year":2025,"lang":"en","type":"article","venue":"SuperIntelligence - Robotics - Safety & Alignment","topic":"Big Data and Business Intelligence","field":"Business, Management and Accounting","cited_by":2,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Institute on Governance","funders":"","keywords":"Frontier; Econometrics; Computer science; Economics; Geography; Archaeology","retraction":null,"screen_n_in":null,"score":{"opus":0.04681276108165858,"gpt":0.3077370296973856,"spread":0.260924268615727,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0005522512,0.0005585831,0.000582519,0.001163295,0.0002529667,0.000417931,0.001122171,0.0002029421,0.00116622],"category_scores_gemma":[0.00008436209,0.0005497142,0.0001930443,0.002524242,0.0001358878,0.00116939,0.0008335178,0.0003456053,0.0007162672],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003602584,"about_ca_system_score_gemma":0.00009431527,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0009054704,"about_ca_topic_score_gemma":0.0009434585,"domain_scores_codex":[0.9964074,0.00001647458,0.000998596,0.0009790622,0.0006402787,0.0009581909],"domain_scores_gemma":[0.9985878,0.00004161118,0.0001171968,0.0009377371,0.0002541399,0.00006145594],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0003254541,0.0005105848,0.01466244,0.0002195729,0.00008253494,0.00003360992,0.0002884226,0.7146886,0.0002939108,0.1421556,0.02541822,0.101321],"study_design_scores_gemma":[0.0004362282,0.00003441627,0.00381861,0.000531668,0.0001261695,0.000004110148,0.0006833831,0.7112631,0.001063851,0.01589065,0.2649875,0.001160369],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.006181432,0.0003691927,0.9418161,0.013768,0.002229226,0.0008269249,0.00007369609,0.0002002974,0.03453512],"genre_scores_gemma":[0.9750209,0.0002393403,0.005648276,0.009851198,0.0004165173,0.0001285249,0.000220095,0.00007323486,0.008401906],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9688395,"threshold_uncertainty_score":0.9997469,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4415757988","doi":"10.70777/si.v2i5.16331","title":"A Grading Rubric for AI Safety Frameworks","year":2025,"lang":"","type":"article","venue":"SuperIntelligence - Robotics - Safety & Alignment","topic":"Ethics and Social Impacts of AI","field":"Social Sciences","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Institute on Governance","funders":"","keywords":"Rubric; Grading (engineering); Warrant; Delphi method; Delphi","retraction":null,"screen_n_in":null,"score":{"opus":0.0323729125467903,"gpt":0.3708621776750627,"spread":0.3384892651282724,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","sts","scholarly_communication","research_integrity","insufficient_payload"],"consensus_categories":["research_integrity"],"category_scores_codex":[0.005357329,0.001116984,0.001514713,0.0003855668,0.005202257,0.001344671,0.002132237,0.002564434,0.0009360673],"category_scores_gemma":[0.00280202,0.001254028,0.001243946,0.002148629,0.002060518,0.0009698975,0.0006891242,0.002822229,0.0001543385],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002537464,"about_ca_system_score_gemma":0.002403239,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001726334,"about_ca_topic_score_gemma":0.0008915103,"domain_scores_codex":[0.9904405,0.0007324198,0.002458148,0.001727639,0.00177436,0.002866951],"domain_scores_gemma":[0.9928394,0.003150532,0.0004562662,0.001248757,0.001314406,0.0009906531],"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":[0.0003935279,0.0006536615,0.0007011672,0.0003861628,0.0007293353,0.00001497408,0.02958393,0.0426941,0.0001157383,0.8947176,0.002545314,0.02746446],"study_design_scores_gemma":[0.001173043,0.0009360185,0.0005132201,0.002605356,0.001037295,0.000003978033,0.0498033,0.02178906,0.001735515,0.5255078,0.3921654,0.002730047],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0005486151,0.003590551,0.820213,0.1267953,0.01003238,0.003795617,0.0002591919,0.0002190277,0.03454631],"genre_scores_gemma":[0.9229221,0.03941546,0.01327055,0.0126065,0.001200078,0.0001405099,0.00006550243,0.0001412426,0.01023806],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9223735,"threshold_uncertainty_score":0.9999772,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4408666991","doi":"10.70777/si.v2i1.13799","title":"Can a Bayesian Oracle Prevent Harm from an Agent?","year":2025,"lang":"en","type":"article","venue":"SuperIntelligence - Robotics - Safety & Alignment","topic":"Blockchain Technology Applications and Security","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal","funders":"","keywords":"Harm; Oracle; Bayesian probability; Computer science; Psychology; Artificial intelligence; Social psychology","retraction":null,"screen_n_in":null,"score":{"opus":0.01427476112546157,"gpt":0.2769327674598674,"spread":0.2626580063344058,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0003597389,0.0003284859,0.0003366052,0.0001594301,0.0004052914,0.0001489979,0.002198328,0.0002146355,0.0001209801],"category_scores_gemma":[0.00002298317,0.0003388515,0.0001277184,0.0007203518,0.0001809,0.0002000797,0.0007515654,0.0003033749,0.00007463865],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003001647,"about_ca_system_score_gemma":0.000186162,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0004480202,"about_ca_topic_score_gemma":0.0003226073,"domain_scores_codex":[0.9973372,0.0001136393,0.0006192104,0.001000076,0.0003589666,0.0005709009],"domain_scores_gemma":[0.9974709,0.00008979029,0.00009632735,0.002031744,0.0000941792,0.0002170404],"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":[0.00001500391,0.0009720396,0.0008907181,0.0000207599,0.0001417062,0.00002746654,0.002425872,0.01337817,0.001320124,0.8442159,0.0003440387,0.1362482],"study_design_scores_gemma":[0.0005406741,0.0004021821,0.004691902,0.0001791315,0.0001204717,0.00001933311,0.001353515,0.3824493,0.1062707,0.4475686,0.05499896,0.001405249],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.006838423,0.0004146214,0.9758202,0.01454792,0.0004965031,0.0006186365,0.00003881851,0.0004316664,0.0007932006],"genre_scores_gemma":[0.9023917,0.0004094417,0.09566878,0.001004349,0.00004261609,0.000118533,0.00002986284,0.00001752539,0.0003172218],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.8955532,"threshold_uncertainty_score":0.9999064,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4408666990","doi":"10.70777/si.v2i1.13975","title":"Effective Mitigations for Systemic Risks from General-Purpose AI","year":2025,"lang":"en","type":"article","venue":"SuperIntelligence - Robotics - Safety & Alignment","topic":"Risk and Safety Analysis","field":"Decision Sciences","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Institute on Governance","funders":"","keywords":"Systemic risk; Computer science; Economics","retraction":null,"screen_n_in":null,"score":{"opus":0.07041623951217256,"gpt":0.4075346826216562,"spread":0.3371184431094836,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.002189176,0.0004730692,0.000938723,0.0004541235,0.0007807153,0.0003766828,0.001463264,0.0002521312,0.000416862],"category_scores_gemma":[0.001142919,0.0003955215,0.0007912266,0.001506298,0.0002739285,0.0003377486,0.000290244,0.000283523,0.0005695809],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004780758,"about_ca_system_score_gemma":0.0002399509,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0008051616,"about_ca_topic_score_gemma":0.0002129641,"domain_scores_codex":[0.9942628,0.0005247755,0.001834205,0.001336214,0.001362653,0.0006793958],"domain_scores_gemma":[0.9936231,0.003472329,0.0002812947,0.001589521,0.0007594042,0.0002743744],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0002595088,0.0002654931,0.01037618,0.00003473516,0.000652473,0.00001787091,0.001415205,0.8221027,0.003837257,0.05934804,0.003647334,0.09804324],"study_design_scores_gemma":[0.001504715,0.0004305123,0.01143313,0.0005176209,0.001207176,0.00002454427,0.007494767,0.2821409,0.0619232,0.5525585,0.0791002,0.001664759],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01756761,0.001407863,0.9687878,0.007796708,0.001168173,0.001799694,0.0003164838,0.00009325221,0.001062346],"genre_scores_gemma":[0.970849,0.001544754,0.02077783,0.001213467,0.0002470868,0.0004029308,0.000132568,0.00003615212,0.004796151],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9532815,"threshold_uncertainty_score":0.9998497,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4417101019","doi":"10.70777/si.v2i4.16671","title":"International AI Safety Report 2025: Second Key Update: Technical Safeguards and Risk Management","year":2025,"lang":"","type":"article","venue":"SuperIntelligence - Robotics - Safety & Alignment","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"University of Toronto; Mila - Quebec Artificial Intelligence Institute","funders":"","keywords":"Key (lock); Risk management; SAFER; Risk assessment; Safeguard; Corporate governance; Risk governance","retraction":null,"screen_n_in":null,"score":{"opus":0.009785592565935903,"gpt":0.2868525432628055,"spread":0.2770669506968696,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","sts","insufficient_payload"],"consensus_categories":["metaepi_narrow"],"category_scores_codex":[0.003977937,0.001388208,0.001322303,0.0007069146,0.001452243,0.00100099,0.004015722,0.0006407808,0.0026646],"category_scores_gemma":[0.0004464759,0.001535504,0.0005437824,0.001555812,0.001008206,0.001180313,0.007950736,0.002242721,0.0002645967],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001952371,"about_ca_system_score_gemma":0.0005244201,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001678956,"about_ca_topic_score_gemma":0.00009254226,"domain_scores_codex":[0.9884675,0.0007138882,0.003545367,0.003414403,0.002157464,0.001701432],"domain_scores_gemma":[0.9944174,0.0004904865,0.0008444645,0.003151747,0.0005027934,0.0005931635],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0004431104,0.0005083493,0.005998868,0.0002778447,0.001414009,0.001258363,0.0007779606,0.4701502,0.00004760301,0.3320492,0.001494022,0.1855805],"study_design_scores_gemma":[0.001654243,0.0003259418,0.01208614,0.001036694,0.0008193096,0.0006527143,0.001011902,0.4330889,0.0008346702,0.01480465,0.531373,0.002311853],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0002391154,0.001197408,0.9278736,0.01265224,0.008797884,0.001640164,0.0001182301,0.000307492,0.04717391],"genre_scores_gemma":[0.4822765,0.04803244,0.4436655,0.002419196,0.0008358224,0.0001538408,0.0002449504,0.0001938775,0.02217783],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.529879,"threshold_uncertainty_score":0.9998868,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4412505613","doi":"10.70777/si.v2i3.15137","title":"Comparing Apples to Oranges: A Taxonomy for Navigating the Global Landscape of AI Regulation","year":2025,"lang":"en","type":"article","venue":"SuperIntelligence - Robotics - Safety & Alignment","topic":"Ethics and Social Impacts of AI","field":"Social Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"","keywords":"Taxonomy (biology); Geography; Biology; Ecology","retraction":null,"screen_n_in":null,"score":{"opus":0.07006106803144764,"gpt":0.3828773494217818,"spread":0.3128162813903342,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001479263,0.0001541068,0.0002840148,0.00002870437,0.0008849181,0.0001574062,0.0004936892,0.0001262061,0.00003525009],"category_scores_gemma":[0.00045887,0.0001334653,0.0001382024,0.0004877765,0.0002920614,0.0001529525,0.0001330882,0.0001936739,0.00000667771],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003113671,"about_ca_system_score_gemma":0.0003636064,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00230825,"about_ca_topic_score_gemma":0.003854942,"domain_scores_codex":[0.998279,0.0001298317,0.0004773385,0.0002639843,0.0004325589,0.0004173223],"domain_scores_gemma":[0.9987134,0.000442259,0.000104426,0.0002661019,0.0003487314,0.0001250416],"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":[0.00006708506,0.00007789544,0.02667503,0.00007033766,0.00009035674,3.677514e-7,0.01368898,0.09936322,0.0001165012,0.8498672,0.0005003444,0.009482646],"study_design_scores_gemma":[0.001352654,0.0006668397,0.02245766,0.002581702,0.0004749757,0.000001884609,0.24777,0.02740169,0.00997241,0.3682527,0.3172857,0.001781884],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03262373,0.0004548286,0.8618183,0.07595792,0.001128822,0.003603445,0.00007627009,0.0001046578,0.02423208],"genre_scores_gemma":[0.9905366,0.0001361219,0.008006253,0.0008995444,0.0001421339,0.00008389985,0.00001115509,0.000008437936,0.000175892],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9579129,"threshold_uncertainty_score":0.6806164,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4415233743","doi":"10.70777/si.v2i6.16169","title":"Responsible Agentic Reasoning and AI Agents: A Critical Survey","year":2025,"lang":"en","type":"article","venue":"SuperIntelligence - Robotics - Safety & Alignment","topic":"Logic, Reasoning, and Knowledge","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Vector Institute","funders":"","keywords":"Operationalization; Intersection (aeronautics); Audit; Trustworthiness; Automated reasoning; Soundness; Class (philosophy)","retraction":null,"screen_n_in":null,"score":{"opus":0.02827062175905466,"gpt":0.3202839762963827,"spread":0.292013354537328,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.001753993,0.0004223152,0.0004860887,0.0002369125,0.0005387939,0.0004823691,0.001199018,0.0001766684,0.000121434],"category_scores_gemma":[0.001263521,0.0004068125,0.0001454772,0.0009092567,0.0003306117,0.0004791987,0.001148003,0.0003457383,0.0001629114],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002687029,"about_ca_system_score_gemma":0.0003890035,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001889031,"about_ca_topic_score_gemma":0.0001258329,"domain_scores_codex":[0.996265,0.0005263743,0.0007080091,0.001081424,0.0005182147,0.0009009768],"domain_scores_gemma":[0.9972624,0.0009239914,0.00006679472,0.001122147,0.0002748262,0.0003498423],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001072032,0.0003711572,0.01693825,0.0001334408,0.0001323134,0.0001193947,0.001920423,0.004643952,0.00031035,0.9616079,0.001734825,0.01198073],"study_design_scores_gemma":[0.001953227,0.001373423,0.186291,0.001880733,0.0003369996,0.0002962219,0.00170756,0.5961468,0.02261992,0.13107,0.05240444,0.003919701],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.003622492,0.00159856,0.9837622,0.002935394,0.001523141,0.0004056073,0.00001284582,0.0002205441,0.005919202],"genre_scores_gemma":[0.9534121,0.001025356,0.04135245,0.001311166,0.00006385487,0.00003207259,0.00001395284,0.00002713667,0.002761896],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9497896,"threshold_uncertainty_score":0.9998384,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4415454748","doi":"10.70777/si.v2i6.16253","title":"International Al Safety Report: First Key Update Capabilities and Risk Implications","year":2025,"lang":"","type":"article","venue":"SuperIntelligence - Robotics - Safety & Alignment","topic":"Aluminum toxicity and tolerance in plants and animals","field":"Agricultural and Biological Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto; Institute on Governance; Mila - Quebec Artificial Intelligence Institute","funders":"European Commission","keywords":"Key (lock); Field (mathematics); Information system; Information technology","retraction":null,"screen_n_in":null,"score":{"opus":0.01627614003752266,"gpt":0.2565377512224817,"spread":0.2402616111849591,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","sts","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0009240177,0.0006265376,0.0006730214,0.00006292728,0.001888448,0.0004365885,0.0008919382,0.0003128943,0.002567199],"category_scores_gemma":[0.000205129,0.0003744373,0.0003295676,0.0004462978,0.000905012,0.000388266,0.0006343348,0.0004950883,0.00009750757],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002750746,"about_ca_system_score_gemma":0.00008710267,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00141789,"about_ca_topic_score_gemma":0.001410075,"domain_scores_codex":[0.99545,0.0001648227,0.001668347,0.001337133,0.0005320247,0.0008477367],"domain_scores_gemma":[0.9980127,0.0006263478,0.0003919193,0.0004094876,0.0002441968,0.0003154064],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"not_applicable","study_design_scores_codex":[0.001386943,0.002463355,0.3780713,0.0003076098,0.002123416,0.0002237323,0.003922388,0.04533178,0.004144053,0.2111272,0.007016246,0.343882],"study_design_scores_gemma":[0.0002858346,0.0002994571,0.1288687,0.0004272369,0.0002243092,0.0001634069,0.003502107,0.004507986,0.001043231,0.008148923,0.8515625,0.0009662932],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5805536,0.009690018,0.05076515,0.2912253,0.01107976,0.004150532,0.0100347,0.0004383301,0.04206267],"genre_scores_gemma":[0.8862237,0.1071195,0.001356428,0.001064412,0.0004017365,0.00004573707,0.0003196134,0.000007277779,0.003461555],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.8445462,"threshold_uncertainty_score":0.9998708,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null}]}