{"paper":{"title":"False Claims, Credible Images: A Red-Teaming Benchmark for Commercial Image Generators","arxiv_id":"2610.11112","primary_cat":"cs.CR","submitted_at":"2026-10-08T02:35:39+00:00","version_compiled":"v1"},"notice":"Every quoted sentence is the authors' text from the arXiv listing abstract. Labels and probabilities are the outputs of classifier-1.13.0 on the question bank abstract-p1-20260923, exactly as returned. No sentence in this record was written by a language model.","schema":"pith.compiled-record/v0","status":"compiled","abstract":{"source":"arXiv listing abstract v1","sentences":[{"i":0,"p":1.0,"id":"s0","own":false,"role":"background","text":"Image-generation models can now produce text-rich, natural-looking visual artifacts that are hard to distinguish from real-world evidence, such as news reports and textbook pages.","hedge":{"p":0.94,"label":"asserted"},"own_p":0.06,"main_claim_p":0.03,"quantitative":false,"quantitative_p":0.11,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.0,"background":1.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":1,"p":0.98,"id":"s1","own":false,"role":"background","text":"Yet, the same capability introduces a new risk: these models can just as easily fabricate visual misinformation.","hedge":{"p":0.69,"label":"asserted"},"own_p":0.08,"main_claim_p":0.09,"quantitative":false,"quantitative_p":0.15,"role_probabilities":{"gap":0.01,"method":0.0,"result":0.0,"background":0.98,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.01}},{"i":2,"p":0.51,"id":"s2","own":false,"role":"result","text":"Even commercial models (e.g., GPT-Image-2) readily produce it.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.41,"main_claim_p":0.05,"quantitative":false,"quantitative_p":0.08,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.51,"background":0.49,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":3,"p":0.95,"id":"s3","own":true,"role":"result","text":"Curiously, we find that these models can recognize a claim as false when asked, yet still render that very claim as credible visual evidence.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.91,"main_claim_p":0.24,"quantitative":false,"quantitative_p":0.11,"role_probabilities":{"gap":0.01,"method":0.0,"result":0.95,"background":0.01,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.03}},{"i":4,"p":0.98,"id":"s4","own":false,"role":"gap","text":"This discrepancy points to a blind spot in current alignment: safeguards judge what an image shows, not what it asserts; however, existing red-teaming benchmarks target conventional harmful content, such as violent or explicit imagery, and say little about where the alignment boundaries lie for visual misinformation, especially in commercial models.","hedge":{"p":0.89,"label":"asserted"},"own_p":0.26,"main_claim_p":0.18,"quantitative":false,"quantitative_p":0.18,"role_probabilities":{"gap":0.98,"method":0.0,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.02}},{"i":5,"p":0.97,"id":"s5","own":true,"role":"contribution","text":"To fill this gap, we introduce EpiReal-Bench, the first systematic benchmark for evaluating visual misinformation risks in commercial image generators, comprising 10k false-claim prompts and 10k corresponding generated images that span 10 real-world claim categories and 10 credible visual formats.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.96,"main_claim_p":0.09,"quantitative":true,"quantitative_p":0.97,"role_probabilities":{"gap":0.0,"method":0.03,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.97,"interpretation":0.0}},{"i":6,"p":0.85,"id":"s6","own":true,"role":"method","text":"We further introduce EpiReal-Attack, a skill-guided black-box optimization framework that uses Pareto-based selection and multimodal feedback to identify commands that bypass alignment safeguards while preserving visual realism, textual legibility, and semantic fidelity.","hedge":{"p":0.99,"label":"asserted"},"own_p":0.93,"main_claim_p":0.05,"quantitative":false,"quantitative_p":0.04,"role_probabilities":{"gap":0.0,"method":0.85,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.15,"interpretation":0.0}},{"i":7,"p":1.0,"id":"s7","own":true,"role":"result","text":"Experiments on four commercial models reveal that more than 70% of false-claim prompts elicit images that faithfully depict the corresponding misinformation, and EpiReal-Attack pushes this rate to 95%.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.96,"main_claim_p":0.38,"quantitative":true,"quantitative_p":0.99,"role_probabilities":{"gap":0.0,"method":0.0,"result":1.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":8,"p":1.0,"id":"s8","own":false,"role":"interpretation","text":"Most worryingly, these models are only a click away, and their outputs are cheap to spread yet hard to disbelieve, leaving this dimension of alignment largely unguarded.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.26,"main_claim_p":0.51,"quantitative":false,"quantitative_p":0.13,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":1.0}}]},"compiled":{"main_claim":{"p":0.77,"text":"Most worryingly, these models are only a click away, and their outputs are cheap to spread yet hard to disbelieve, leaving this dimension of alignment largely unguarded.","location":"abstract s8","hedge_strength":{"p":1.0,"label":"asserted"}},"provenance":{"route":{"via":"TypeSafe API","model":"classifier-1.13.0","endpoint":"https://api.typesafe.ai/v1/systemone","provider":"TypeSafe"},"packing":"one request per paper","requests":1,"input_tokens":7159,"cost_usd_reported":0.0},"compiled_at":"2026-10-09","bank_version":"abstract-p1-20260923","model_version":"classifier-1.13.0","abstract_states":{"claims_first":0.94,"states_limitation":0.21,"human_participants":0.07,"controlled_comparison":0.21,"randomized_assignment":0.01,"states_quantitative_result":0.99,"compares_with_prior_results":0.68,"mentions_code_or_data_availability":0.05},"evidential_basis":{"p":1.0,"value":"observation or experiment","probabilities":{"argument":0.0,"literature":0.0,"proof_or_derivation":0.0,"analysis_of_existing_data":0.0,"observation_or_experiment":1.0,"simulation_or_computation":0.0}},"contribution_type":{"p":1.0,"value":"dataset or benchmark","probabilities":{"review":0.0,"method_or_tool":0.0,"empirical_study":0.0,"simulation_study":0.0,"theoretical_result":0.0,"dataset_or_benchmark":1.0,"position_or_commentary":0.0}},"contribution_statement":{"p":1.0,"text":"To fill this gap, we introduce EpiReal-Bench, the first systematic benchmark for evaluating visual misinformation risks in commercial image generators, comprising 10k false-claim prompts and 10k corresponding generated images that span 10 real-world claim categories and 10 credible visual formats.","location":"abstract s5"}},"coverage":{"tier":"abstract","levels_run":["P abstract"],"not_examined":["full text","figures and tables","references and citing sentences","relations to other papers"],"sources_read":["arXiv listing abstract v1"],"levels_not_run":["S sections","T body sentences","C citations","R cross-paper relations"],"sentences_examined":9,"sentences_in_abstract":9},"measured_accuracy":{"url":"https://pith.science/accuracy","measured_on":"2026-10-02","method":"two independent model families labelled a fresh random sample blind; the figure is how often the published label matches their consensus","sentence_role_agreement":0.92,"main_claim_agreement":0.908,"typed_citation_agreement":0.933,"typed_citation_agreement_95_interval":[0.917,0.945],"review_comments_accurate":0.758,"review_comments_accurate_95_interval":[0.684,0.82],"review_caught_admitted_error_found_or_related":0.448,"review_comments_judged":{"judged_by":["gpt-5.6-luna","grok-4.7"],"judged_against":"Pith's own text copy of each paper: text extracted from the PDF, the first 90,000 characters","measured_on":"2026-10-04","accurate":113,"decided":149,"sampled_reviews":200,"split_by_lane":false,"by_lane":"https://pith.science/accuracy#by-lane"}}}