{"id":"https://openalex.org/W7204546358","doi":"https://doi.org/10.48550/arxiv.2608.26535","title":"Multi2AV-Safety: Benchmarking Safety in Multimodal-to-Audio-Video Generation","display_name":"Multi2AV-Safety: Benchmarking Safety in Multimodal-to-Audio-Video Generation","publication_year":2026,"publication_date":"2026-08-27","ids":{"openalex":"https://openalex.org/W7204546358","doi":"https://doi.org/10.48550/arxiv.2608.26535"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2608.26535","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.26535","pdf_url":"https://arxiv.org/pdf/2608.26535","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2608.26535","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5148544857","display_name":"Kaichao Jiang","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jiang, Kaichao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5077867418","display_name":"Changtao Miao","orcid":"https://orcid.org/0000-0002-7634-9992","observed_orcids":["https://orcid.org/0000-0002-7634-9992"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Miao, Changtao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148522745","display_name":"Baiqi Wu","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wu, Baiqi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101559520","display_name":"Zhiyuan Lu","orcid":"https://orcid.org/0000-0001-7329-5477","observed_orcids":["https://orcid.org/0000-0001-7329-5477"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lu, Zhiyuan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148524242","display_name":"Kang Yang","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yang, Kang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148519432","display_name":"Peiwei Zhao","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhao, Peiwei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137659253","display_name":"Junchi Chen","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Junchi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5026338785","display_name":"Yunfeng Diao","orcid":"https://orcid.org/0000-0002-9455-1510","observed_orcids":["https://orcid.org/0000-0002-9455-1510"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Diao, Yunfeng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148500117","display_name":"He Liu","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, He","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136401136","display_name":"Qi Chu","orcid":"https://orcid.org/0000-0003-3028-0755","observed_orcids":["https://orcid.org/0000-0003-3028-0755"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chu, Qi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5148482930","display_name":"Tao Gong","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gong, Tao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.9086999893188477,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.9086999893188477,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T12262","display_name":"Hate Speech and Cyberbullying Detection","score":0.05040000006556511,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.013299999758601189,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/safety-benchmark","display_name":"safety benchmark","score":0.578000009059906},{"id":"https://openalex.org/keywords/multimodal-conditioning","display_name":"multimodal conditioning","score":0.8880000114440918},{"id":"https://openalex.org/keywords/harmful-content-detection","display_name":"harmful content detection","score":0.6079999804496765},{"id":"https://openalex.org/keywords/safety-guards","display_name":"safety guards","score":0.5509999990463257}],"concepts":[{"id":"https://openalex.org/C86251818","wikidata":"https://www.wikidata.org/wiki/Q816754","display_name":"Benchmarking","level":2,"score":0.654699981212616},{"id":"https://openalex.org/C2779903281","wikidata":"https://www.wikidata.org/wiki/Q6888026","display_name":"Modalities","level":2,"score":0.6252999901771545},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5855000019073486},{"id":"https://openalex.org/C112930515","wikidata":"https://www.wikidata.org/wiki/Q4389547","display_name":"Risk analysis (engineering)","level":1,"score":0.582099974155426},{"id":"https://openalex.org/C2776743756","wikidata":"https://www.wikidata.org/wiki/Q5097921","display_name":"Safeguarding","level":2,"score":0.5683000087738037},{"id":"https://openalex.org/C2780428219","wikidata":"https://www.wikidata.org/wiki/Q16952335","display_name":"Cover (algebra)","level":2,"score":0.5016000270843506},{"id":"https://openalex.org/C184337299","wikidata":"https://www.wikidata.org/wiki/Q1437428","display_name":"Semantics (computer science)","level":2,"score":0.4275999963283539},{"id":"https://openalex.org/C100776233","wikidata":"https://www.wikidata.org/wiki/Q2532492","display_name":"Bridge (graph theory)","level":2,"score":0.40540000796318054},{"id":"https://openalex.org/C184356942","wikidata":"https://www.wikidata.org/wiki/Q830382","display_name":"Best practice","level":2,"score":0.3849000036716461},{"id":"https://openalex.org/C2780234205","wikidata":"https://www.wikidata.org/wiki/Q2281276","display_name":"Safety case","level":2,"score":0.36340001225471497},{"id":"https://openalex.org/C2780226545","wikidata":"https://www.wikidata.org/wiki/Q6888030","display_name":"Modality (human\u2013computer interaction)","level":2,"score":0.36039999127388},{"id":"https://openalex.org/C132835097","wikidata":"https://www.wikidata.org/wiki/Q7663745","display_name":"System safety","level":2,"score":0.3508000075817108},{"id":"https://openalex.org/C148493468","wikidata":"https://www.wikidata.org/wiki/Q2646951","display_name":"Functional safety","level":2,"score":0.3488999903202057},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.3456000089645386},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.3221000134944916},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.31349998712539673},{"id":"https://openalex.org/C2779119184","wikidata":"https://www.wikidata.org/wiki/Q294350","display_name":"ALARM","level":2,"score":0.2806999981403351},{"id":"https://openalex.org/C49261128","wikidata":"https://www.wikidata.org/wiki/Q1132455","display_name":"Hazard","level":2,"score":0.26919999718666077},{"id":"https://openalex.org/C2779328685","wikidata":"https://www.wikidata.org/wiki/Q1475557","display_name":"Patient safety","level":3,"score":0.2639999985694885},{"id":"https://openalex.org/C43540301","wikidata":"https://www.wikidata.org/wiki/Q689971","display_name":"Paradigm shift","level":2,"score":0.25929999351501465},{"id":"https://openalex.org/C2776654903","wikidata":"https://www.wikidata.org/wiki/Q2601463","display_name":"SAFER","level":2,"score":0.2578999996185303},{"id":"https://openalex.org/C12174686","wikidata":"https://www.wikidata.org/wiki/Q1058438","display_name":"Risk assessment","level":2,"score":0.2533999979496002}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2608.26535","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.26535","pdf_url":"https://arxiv.org/pdf/2608.26535","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2608.26535","doi":"https://doi.org/10.48550/arxiv.2608.26535","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2608.26535","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2608.26535","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.26535","pdf_url":"https://arxiv.org/pdf/2608.26535","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7204546358.pdf","grobid_xml":"https://content.openalex.org/works/W7204546358.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7106692253","https://openalex.org/W7211990609","https://openalex.org/W4416245110","https://openalex.org/W7127575289","https://openalex.org/W4417004297","https://openalex.org/W7211993202","https://openalex.org/W7151242312","https://openalex.org/W4416933465","https://openalex.org/W7161767334","https://openalex.org/W7220446104"],"abstract_inverted_index":{"Recent":[0],"audio-video":[1,32,98],"generators":[2,99],"increasingly":[3],"support":[4],"joint":[5],"conditioning":[6,80],"on":[7,179,202,220],"text,":[8],"images,":[9],"audio,":[10],"and":[11,29,52,85,100,110,116,137,150,222],"video.":[12],"These":[13],"capabilities":[14],"also":[15],"enable":[16,159],"attacks":[17,201],"that":[18,158],"exploit":[19],"cross-modal":[20],"interactions":[21],"or":[22],"obscure":[23],"harmful":[24,31],"intent":[25],"to":[26,163,176],"bypass":[27],"safeguards":[28,198],"induce":[30],"outputs.":[33],"However,":[34],"existing":[35],"generation-safety":[36],"benchmarks":[37],"have":[38],"not":[39],"kept":[40],"pace":[41],"with":[42,112,191],"these":[43,58,127],"advances,":[44],"providing":[45],"limited":[46],"coverage":[47],"of":[48,91,218],"multimodal":[49,113,188,200],"input":[50],"combinations":[51],"obscured":[53,117],"attack":[54,73],"intents.":[55],"To":[56],"address":[57],"gaps,":[59],"we":[60,129],"introduce":[61,130],"Multi2AV-Safety,":[62],"the":[63,174,211],"first":[64],"full-coverage":[65],"red-team":[66],"benchmark":[67],"for":[68],"multimodal-to-audio-video":[69],"generation,":[70],"comprising":[71],"11,024":[72],"instances":[74],"across":[75],"all":[76],"11":[77],"non-singleton":[78],"T/I/A/V":[79],"configurations,":[81],"4":[82],"attack-intent":[83,118,138],"categories,":[84],"5":[86],"harm":[87],"categories.":[88],"Our":[89],"evaluation":[90],"recent":[92],"state-of-the-art":[93],"models,":[94],"including":[95],"four":[96],"multimodal-conditioned":[97],"eight":[101],"safety":[102,151,161,183,189],"guards,":[103],"reveals":[104],"substantial":[105],"vulnerabilities":[106],"in":[107],"both":[108],"generation":[109,149],"safeguarding,":[111],"compositional":[114],"risk":[115,119,142,156],"emerging":[120],"as":[121],"two":[122],"complementary":[123],"challenges.":[124],"Guided":[125],"by":[126,225],"findings,":[128],"PerceptGuard,":[131],"an":[132,215],"omni-modal":[133],"guard":[134],"integrating":[135],"compositional-risk":[136],"supervision":[139],"through":[140],"structured":[141],"perception":[143],"learning.":[144],"By":[145],"jointly":[146],"training":[147],"rationale":[148,170],"classification,":[152],"it":[153,207],"learns":[154],"shared":[155],"representations":[157],"a":[160],"head":[162],"make":[164],"efficient":[165],"predictions":[166],"at":[167],"inference":[168],"without":[169],"decoding,":[171],"while":[172],"retaining":[173],"ability":[175],"generate":[177],"explanations":[178],"demand.":[180],"Across":[181],"34":[182],"benchmarks,":[184],"PerceptGuard":[185],"combines":[186],"SOTA":[187],"detection":[190],"highly":[192],"competitive":[193],"unimodal":[194],"performance,":[195],"strengthening":[196],"input-side":[197],"against":[199,210],"omni":[203],"models.":[204],"In":[205],"particular,":[206],"improves":[208],"safeguarding":[209],"above":[212],"risks,":[213],"achieving":[214],"overall":[216],"recall":[217],"86.06\\%":[219],"Multi2AV-Safety":[221],"outperforming":[223],"GuardReasoner-Omni":[224],"14.56\\%.":[226]},"counts_by_year":[],"updated_date":"2026-10-28T10:33:24.094632","created_date":"2026-08-29T00:00:00"}