{"id":"https://openalex.org/W4416036395","doi":"https://doi.org/10.18653/v1/2025.emnlp-main.804","title":"Your RAG is Unfair: Exposing Fairness Vulnerabilities in Retrieval-Augmented Generation via Backdoor Attacks","display_name":"Your RAG is Unfair: Exposing Fairness Vulnerabilities in Retrieval-Augmented Generation via Backdoor Attacks","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4416036395","doi":"https://doi.org/10.18653/v1/2025.emnlp-main.804"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2025.emnlp-main.804","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.emnlp-main.804","pdf_url":"https://aclanthology.org/2025.emnlp-main.804.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2025.emnlp-main.804.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5073718577","display_name":"Gaurav Bagwe","orcid":"https://orcid.org/0000-0001-5706-5065"},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Gaurav Bagwe","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5057682656","display_name":"Saket S. Chaturvedi","orcid":"https://orcid.org/0000-0003-0700-404X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Saket Sanjeev Chaturvedi","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5115602381","display_name":"Xiaolong Ma","orcid":"https://orcid.org/0000-0002-3151-4261"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xiaolong Ma","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5010643450","display_name":"Xiaoyong Yuan","orcid":"https://orcid.org/0000-0003-0782-4187"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xiaoyong Yuan","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5112529490","display_name":"Kuang-Ching Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kuang-Ching Wang","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5100322310","display_name":"Lan Zhang","orcid":"https://orcid.org/0000-0002-7718-6128"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lan Emily Zhang","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":6,"corresponding_author_ids":["https://openalex.org/A5073718577"],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.1750949,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"15930","last_page":"15948"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.16089999675750732,"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":0.16089999675750732,"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/T11644","display_name":"Spam and Phishing Detection","score":0.11550000309944153,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/T12479","display_name":"Web Application Security Vulnerabilities","score":0.09200000017881393,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/backdoor","display_name":"Backdoor","score":0.8970999717712402},{"id":"https://openalex.org/keywords/vulnerability","display_name":"Vulnerability (computing)","score":0.30640000104904175},{"id":"https://openalex.org/keywords/culpability","display_name":"Culpability","score":0.27970001101493835},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.2667999863624573}],"concepts":[{"id":"https://openalex.org/C2781045450","wikidata":"https://www.wikidata.org/wiki/Q254569","display_name":"Backdoor","level":2,"score":0.8970999717712402},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.6801000237464905},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5582000017166138},{"id":"https://openalex.org/C95713431","wikidata":"https://www.wikidata.org/wiki/Q631425","display_name":"Vulnerability (computing)","level":2,"score":0.30640000104904175},{"id":"https://openalex.org/C144133560","wikidata":"https://www.wikidata.org/wiki/Q4830453","display_name":"Business","level":0,"score":0.2946999967098236},{"id":"https://openalex.org/C2779990267","wikidata":"https://www.wikidata.org/wiki/Q1365664","display_name":"Culpability","level":2,"score":0.27970001101493835},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.2667999863624573},{"id":"https://openalex.org/C178005623","wikidata":"https://www.wikidata.org/wiki/Q308859","display_name":"Anonymity","level":2,"score":0.24060000479221344},{"id":"https://openalex.org/C108827166","wikidata":"https://www.wikidata.org/wiki/Q175975","display_name":"Internet privacy","level":1,"score":0.23909999430179596},{"id":"https://openalex.org/C31258907","wikidata":"https://www.wikidata.org/wiki/Q1301371","display_name":"Computer network","level":1,"score":0.2257000058889389}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2025.emnlp-main.804","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.emnlp-main.804","pdf_url":"https://aclanthology.org/2025.emnlp-main.804.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.emnlp-main.804","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.emnlp-main.804","pdf_url":"https://aclanthology.org/2025.emnlp-main.804.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G1068985279","display_name":"CloudLab Phase-IV: Strengthening a Core Resource for Research and Education","funder_award_id":"2431419","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"},{"id":"https://openalex.org/G6323360741","display_name":null,"funder_award_id":"2330891","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"},{"id":"https://openalex.org/G848032724","display_name":null,"funder_award_id":"Science","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"},{"id":"https://openalex.org/G8697316022","display_name":"Collaborative Research: SHF: Medium: Heterogeneous Architecture for Collaborative Machine Learning","funder_award_id":"2426318","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"}],"funders":[{"id":"https://openalex.org/F4320306076","display_name":"National Science Foundation","ror":"https://ror.org/021nxhr62"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4416036395.pdf","grobid_xml":"https://content.openalex.org/works/W4416036395.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Retrieval-augmented":[0],"generation":[1,50],"(RAG)":[2],"enhances":[3],"factual":[4],"grounding":[5],"by":[6],"integrating":[7],"retrieval":[8,48],"mechanisms":[9],"with":[10,101],"generative":[11],"models":[12],"but":[13],"introduces":[14,72],"new":[15],"attack":[16,152],"surfaces,":[17],"particularly":[18],"through":[19,83],"backdoor":[20,86],"attacks.While":[21],"prior":[22],"research":[23],"has":[24],"largely":[25],"focused":[26],"on":[27,38,68],"disinformation":[28],"threats,":[29],"fairness":[30,79,132,168],"vulnerabilities":[31,80],"remain":[32],"underexplored.Unlike":[33],"conventional":[34],"backdoors":[35],"that":[36,77,148],"rely":[37],"direct":[39],"trigger-to-target":[40],"mappings,":[41],"fairness-driven":[42],"attacks":[43],"exploit":[44],"the":[45,88,91,98,102,109,121],"interaction":[46],"between":[47,55],"and":[49,58,65,144,159,164],"models,":[51],"manipulating":[52],"semantic":[53],"relationships":[54],"target":[56,99,137],"groups":[57],"social":[59,104],"biases":[60],"to":[61,96,119,167],"establish":[62],"a":[63,74,84,162],"persistent":[64,163],"covert":[66],"influence":[67],"content":[69,126],"generation.This":[70],"paper":[71],"BiasRAG,":[73],"systematic":[75],"framework":[76],"exposes":[78],"in":[81,169],"RAG":[82],"two-phase":[85],"attack.During":[87],"pre-training":[89],"phase,":[90,111],"query":[92],"encoder":[93],"is":[94],"compromised":[95],"align":[97],"group":[100],"intended":[103],"bias,":[105],"ensuring":[106],"longterm":[107],"persistence.In":[108],"post-deployment":[110],"adversarial":[112],"documents":[113],"are":[114],"injected":[115],"into":[116],"knowledge":[117],"bases":[118],"reinforce":[120],"backdoor,":[122],"subtly":[123],"influencing":[124],"retrieved":[125],"while":[127,155],"remaining":[128],"undetectable":[129],"under":[130],"standard":[131],"evaluations.Together,":[133],"BiasRAG":[134,149],"ensures":[135],"precise":[136],"alignment":[138],"over":[139],"sensitive":[140],"attributes,":[141],"stealthy":[142],"execution,":[143],"resilience.Empirical":[145],"evaluations":[146],"demonstrate":[147],"achieves":[150],"high":[151],"success":[153],"rates":[154],"preserving":[156],"contextual":[157],"relevance":[158],"utility,":[160],"establishing":[161],"evolving":[165],"threat":[166],"RAG.":[170]},"counts_by_year":[],"updated_date":"2026-04-10T15:06:20.359241","created_date":"2025-11-08T00:00:00"}