{"id":"W4413212461","doi":"10.1109/ms.2025.3597574","title":"When AI-Generated Unit Tests Validate Bugs: The Risk of Faulty Assertions","year":2025,"lang":"en","type":"article","venue":"IEEE Software","topic":"Software Testing and Debugging Techniques","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Software bug; Computer science; Unit testing; Software engineering; Programming language; Reliability engineering; Software; 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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000562797,0.0001878946,0.0002131987,0.0001674755,0.000531868,0.0001272583,0.001373723,0.0001070284,0.000007057709],"category_scores_gemma":[0.001676502,0.0001358168,0.0001071266,0.000999197,0.0001056203,0.0002092676,0.0003152339,0.0003180067,0.00001763212],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003147524,"about_ca_system_score_gemma":0.0002162318,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0005520236,"about_ca_topic_score_gemma":0.00004764487,"domain_scores_codex":[0.9985343,0.0002278015,0.000327551,0.0003699576,0.0002605915,0.0002797912],"domain_scores_gemma":[0.9970469,0.001040392,0.000171455,0.001243333,0.0004412721,0.00005659091],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.00001815245,0.0003537072,0.4749056,0.0001196269,0.0003076962,0.0000248961,0.002616516,0.002738651,0.001272636,0.005041053,0.4135978,0.09900373],"study_design_scores_gemma":[0.001132633,0.0003354127,0.156303,0.000914842,0.0002850448,0.00004218929,0.0000340365,0.05373204,0.06803026,0.6923715,0.02564053,0.001178463],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.06111348,0.0002107561,0.9280491,0.001275041,0.0005356716,0.0002206622,0.00003465508,0.008374388,0.0001862465],"genre_scores_gemma":[0.807926,0.00002151597,0.1900817,0.001104803,0.00004653792,0.00005701689,0.00000942226,0.00001638249,0.0007365832],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.7468125,"threshold_uncertainty_score":0.5538447,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02597969980246511,"score_gpt":0.2993781457658547,"score_spread":0.2733984459633896,"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."}}