{"id":"https://openalex.org/W7140646741","doi":"https://doi.org/10.48550/arxiv.2603.24079","title":"When Understanding Becomes a Risk: Authenticity and Safety Risks in the Emerging Image Generation Paradigm","display_name":"When Understanding Becomes a Risk: Authenticity and Safety Risks in the Emerging Image Generation Paradigm","publication_year":2026,"publication_date":"2026-03-25","ids":{"openalex":"https://openalex.org/W7140646741","doi":"https://doi.org/10.48550/arxiv.2603.24079"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2603.24079","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.24079","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":"article"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2603.24079","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5130709472","display_name":"Ye Leng","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Leng, Ye","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5106708808","display_name":"Junjie Chu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chu, Junjie","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130712203","display_name":"Mingjie Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Mingjie","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130664332","display_name":"Chenhao Lin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lin, Chenhao","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130694112","display_name":"Chao Shen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shen, Chao","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130646577","display_name":"Michael Backes","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Backes, Michael","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130676991","display_name":"Yun Shen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shen, Yun","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5130640590","display_name":"Yang Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Yang","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":8,"corresponding_author_ids":["https://openalex.org/A5130709472"],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.26030001044273376,"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"}},"topics":[{"id":"https://openalex.org/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.26030001044273376,"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"}},{"id":"https://openalex.org/T11147","display_name":"Misinformation and Its Impacts","score":0.08250000327825546,"subfield":{"id":"https://openalex.org/subfields/3312","display_name":"Sociology and Political Science"},"field":{"id":"https://openalex.org/fields/33","display_name":"Social Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.07900000363588333,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/process","display_name":"Process (computing)","score":0.5551999807357788},{"id":"https://openalex.org/keywords/image","display_name":"Image (mathematics)","score":0.5044000148773193},{"id":"https://openalex.org/keywords/generative-grammar","display_name":"Generative grammar","score":0.48910000920295715},{"id":"https://openalex.org/keywords/generative-model","display_name":"Generative model","score":0.3871999979019165},{"id":"https://openalex.org/keywords/benchmark","display_name":"Benchmark (surveying)","score":0.38089999556541443},{"id":"https://openalex.org/keywords/semantics","display_name":"Semantics (computer science)","score":0.33239999413490295}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6330000162124634},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.5551999807357788},{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"score":0.5044000148773193},{"id":"https://openalex.org/C39890363","wikidata":"https://www.wikidata.org/wiki/Q36108","display_name":"Generative grammar","level":2,"score":0.48910000920295715},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4205000102519989},{"id":"https://openalex.org/C167966045","wikidata":"https://www.wikidata.org/wiki/Q5532625","display_name":"Generative model","level":3,"score":0.3871999979019165},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.38089999556541443},{"id":"https://openalex.org/C112930515","wikidata":"https://www.wikidata.org/wiki/Q4389547","display_name":"Risk analysis (engineering)","level":1,"score":0.35359999537467957},{"id":"https://openalex.org/C184337299","wikidata":"https://www.wikidata.org/wiki/Q1437428","display_name":"Semantics (computer science)","level":2,"score":0.33239999413490295},{"id":"https://openalex.org/C2522767166","wikidata":"https://www.wikidata.org/wiki/Q2374463","display_name":"Data science","level":1,"score":0.3131999969482422},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.31049999594688416},{"id":"https://openalex.org/C2780554381","wikidata":"https://www.wikidata.org/wiki/Q2063340","display_name":"Sensemaking","level":2,"score":0.26750001311302185},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.2637999951839447},{"id":"https://openalex.org/C90312973","wikidata":"https://www.wikidata.org/wiki/Q7449052","display_name":"Semantic data model","level":2,"score":0.2635999917984009},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.26159998774528503},{"id":"https://openalex.org/C43540301","wikidata":"https://www.wikidata.org/wiki/Q689971","display_name":"Paradigm shift","level":2,"score":0.2547000050544739}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2603.24079","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.24079","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":null,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2603.24079","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.24079","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":"article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Recently,":[0],"multimodal":[1],"large":[2],"language":[3,13],"models":[4,59,110],"(MLLMs)":[5],"have":[6,181],"emerged":[7],"as":[8,60],"a":[9,23,61],"unified":[10],"paradigm":[11],"for":[12,27],"and":[14,38,52,67,81,126,163],"image":[15,83,134],"generation.":[16],"Compared":[17],"with":[18,149,161],"diffusion":[19,58,102,109],"models,":[20],"MLLMs":[21,74,94,121,160],"possess":[22],"much":[24],"stronger":[25],"capability":[26],"semantic":[28,46],"understanding,":[29],"enabling":[30],"them":[31],"to":[32,96,113,142,189],"process":[33],"more":[34,98,164],"complex":[35],"textual":[36],"inputs":[37],"comprehend":[39,123],"richer":[40],"contextual":[41],"meanings.":[42],"However,":[43],"this":[44],"enhanced":[45],"ability":[47],"may":[48],"also":[49,139],"introduce":[50],"new":[51,187],"potentially":[53],"greater":[54],"safety":[55,70,173],"risks.":[56],"Taking":[57],"reference":[62],"point,":[63],"we":[64,91],"systematically":[65],"analyze":[66],"compare":[68],"the":[69,171,176],"risks":[71,174],"of":[72,175],"emerging":[73,172],"along":[75],"two":[76],"dimensions:":[77],"unsafe":[78,87,99,128],"content":[79],"generation":[80,88],"fake":[82,133],"synthesis.":[84],"Across":[85],"multiple":[86],"benchmark":[89],"datasets,":[90],"observe":[92],"that":[93,170],"tend":[95],"generate":[97,127],"images":[100,137],"than":[101],"models.":[103],"This":[104],"difference":[105],"partly":[106],"arises":[107],"because":[108],"often":[111],"fail":[112],"interpret":[114],"abstract":[115],"prompts,":[116],"producing":[117],"corrupted":[118],"outputs,":[119],"whereas":[120],"can":[122,153],"these":[124],"prompts":[125],"content.":[129],"For":[130],"current":[131],"advanced":[132],"detectors,":[135],"MLLM-generated":[136],"are":[138,147],"notably":[140],"harder":[141],"identify.":[143],"Even":[144],"when":[145],"detectors":[146],"retrained":[148],"MLLMs-specific":[150],"data,":[151],"they":[152],"still":[154],"be":[155],"bypassed":[156],"by":[157],"simply":[158],"providing":[159],"longer":[162],"descriptive":[165],"inputs.":[166],"Our":[167],"measurements":[168],"indicate":[169],"cutting-edge":[177],"generative":[178],"paradigm,":[179],"MLLMs,":[180],"not":[182],"been":[183],"sufficiently":[184],"recognized,":[185],"posing":[186],"challenges":[188],"real-world":[190],"safety.":[191]},"counts_by_year":[],"updated_date":"2026-03-27T06:05:27.210665","created_date":"2026-03-27T00:00:00"}