{"id":"https://openalex.org/W7140289118","doi":"https://doi.org/10.48550/arxiv.2603.23146","title":"Why AI-Generated Text Detection Fails: Evidence from Explainable AI Beyond Benchmark Accuracy","display_name":"Why AI-Generated Text Detection Fails: Evidence from Explainable AI Beyond Benchmark Accuracy","publication_year":2026,"publication_date":"2026-03-24","ids":{"openalex":"https://openalex.org/W7140289118","doi":"https://doi.org/10.48550/arxiv.2603.23146"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2603.23146","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.23146","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.2603.23146","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5056216309","display_name":"Shushanta Pudasaini","orcid":"https://orcid.org/0000-0003-1612-4510"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Pudasaini, Shushanta","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130325925","display_name":"Luis Miralles-Pechu\u00e1n","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Miralles-Pechu\u00e1n, Luis","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5037769267","display_name":"David Lillis","orcid":"https://orcid.org/0000-0002-5702-4463"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lillis, David","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5075843624","display_name":"Marisa Llorens Salvador","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Salvador, Marisa Llorens","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":4,"corresponding_author_ids":[],"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/T12380","display_name":"Authorship Attribution and Profiling","score":0.7415000200271606,"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/T12380","display_name":"Authorship Attribution and Profiling","score":0.7415000200271606,"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/T10028","display_name":"Topic Modeling","score":0.07079999893903732,"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/T11636","display_name":"Artificial Intelligence in Healthcare and Education","score":0.03290000185370445,"subfield":{"id":"https://openalex.org/subfields/2718","display_name":"Health Informatics"},"field":{"id":"https://openalex.org/fields/27","display_name":"Medicine"},"domain":{"id":"https://openalex.org/domains/4","display_name":"Health Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/interpretability","display_name":"Interpretability","score":0.8543000221252441},{"id":"https://openalex.org/keywords/discriminative-model","display_name":"Discriminative model","score":0.5615000128746033},{"id":"https://openalex.org/keywords/exploit","display_name":"Exploit","score":0.5507000088691711},{"id":"https://openalex.org/keywords/benchmark","display_name":"Benchmark (surveying)","score":0.5286999940872192},{"id":"https://openalex.org/keywords/software-portability","display_name":"Software portability","score":0.43470001220703125},{"id":"https://openalex.org/keywords/feature-engineering","display_name":"Feature engineering","score":0.4165000021457672},{"id":"https://openalex.org/keywords/domain","display_name":"Domain (mathematical analysis)","score":0.3937999904155731},{"id":"https://openalex.org/keywords/clef","display_name":"Clef","score":0.3869999945163727},{"id":"https://openalex.org/keywords/python","display_name":"Python (programming language)","score":0.3862000107765198}],"concepts":[{"id":"https://openalex.org/C2781067378","wikidata":"https://www.wikidata.org/wiki/Q17027399","display_name":"Interpretability","level":2,"score":0.8543000221252441},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7684000134468079},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.6837000250816345},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.5659999847412109},{"id":"https://openalex.org/C97931131","wikidata":"https://www.wikidata.org/wiki/Q5282087","display_name":"Discriminative model","level":2,"score":0.5615000128746033},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.5507000088691711},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5436000227928162},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.5286999940872192},{"id":"https://openalex.org/C63000827","wikidata":"https://www.wikidata.org/wiki/Q3080428","display_name":"Software portability","level":2,"score":0.43470001220703125},{"id":"https://openalex.org/C2778827112","wikidata":"https://www.wikidata.org/wiki/Q22245680","display_name":"Feature engineering","level":3,"score":0.4165000021457672},{"id":"https://openalex.org/C36503486","wikidata":"https://www.wikidata.org/wiki/Q11235244","display_name":"Domain (mathematical analysis)","level":2,"score":0.3937999904155731},{"id":"https://openalex.org/C107763842","wikidata":"https://www.wikidata.org/wiki/Q181040","display_name":"Clef","level":3,"score":0.3869999945163727},{"id":"https://openalex.org/C519991488","wikidata":"https://www.wikidata.org/wiki/Q28865","display_name":"Python (programming language)","level":2,"score":0.3862000107765198},{"id":"https://openalex.org/C176982825","wikidata":"https://www.wikidata.org/wiki/Q835922","display_name":"Lexical analysis","level":2,"score":0.3765999972820282},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.3682999908924103},{"id":"https://openalex.org/C79581498","wikidata":"https://www.wikidata.org/wiki/Q1367530","display_name":"Suite","level":2,"score":0.34200000762939453},{"id":"https://openalex.org/C95623464","wikidata":"https://www.wikidata.org/wiki/Q1096149","display_name":"Classifier (UML)","level":2,"score":0.32820001244544983},{"id":"https://openalex.org/C155092808","wikidata":"https://www.wikidata.org/wiki/Q182557","display_name":"Computational linguistics","level":2,"score":0.32420000433921814},{"id":"https://openalex.org/C2776145971","wikidata":"https://www.wikidata.org/wiki/Q30673951","display_name":"Labeled data","level":2,"score":0.31859999895095825},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.30390000343322754},{"id":"https://openalex.org/C75608658","wikidata":"https://www.wikidata.org/wiki/Q44395","display_name":"Pascal (unit)","level":2,"score":0.3034999966621399},{"id":"https://openalex.org/C23123220","wikidata":"https://www.wikidata.org/wiki/Q816826","display_name":"Information retrieval","level":1,"score":0.3021000027656555},{"id":"https://openalex.org/C2777601683","wikidata":"https://www.wikidata.org/wiki/Q6499736","display_name":"Vocabulary","level":2,"score":0.30160000920295715},{"id":"https://openalex.org/C43214815","wikidata":"https://www.wikidata.org/wiki/Q7310987","display_name":"Reliability (semiconductor)","level":3,"score":0.28630000352859497},{"id":"https://openalex.org/C94915269","wikidata":"https://www.wikidata.org/wiki/Q1834857","display_name":"Detector","level":2,"score":0.28380000591278076},{"id":"https://openalex.org/C52622490","wikidata":"https://www.wikidata.org/wiki/Q1026626","display_name":"Feature extraction","level":2,"score":0.2752000093460083},{"id":"https://openalex.org/C207685749","wikidata":"https://www.wikidata.org/wiki/Q2088941","display_name":"Domain knowledge","level":2,"score":0.2727999985218048},{"id":"https://openalex.org/C46686674","wikidata":"https://www.wikidata.org/wiki/Q466303","display_name":"Boosting (machine learning)","level":2,"score":0.26750001311302185},{"id":"https://openalex.org/C12267149","wikidata":"https://www.wikidata.org/wiki/Q282453","display_name":"Support vector machine","level":2,"score":0.25290000438690186}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2603.23146","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.23146","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.2603.23146","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.23146","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":[{"display_name":"Reduced inequalities","score":0.6922030448913574,"id":"https://metadata.un.org/sdg/10"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"The":[0],"widespread":[1],"adoption":[2],"of":[3,12,103,152],"Large":[4],"Language":[5],"Models":[6],"(LLMs)":[7],"has":[8],"made":[9],"the":[10,131,170,181],"detection":[11,22,62],"AI-Generated":[13],"text":[14,168],"a":[15,162],"pressing":[16],"and":[17,35,71,86,108,190,211,224],"complex":[18],"challenge.":[19],"Although":[20],"many":[21],"systems":[23],"report":[24],"high":[25],"benchmark":[26,80],"accuracy,":[27],"their":[28,36],"reliability":[29],"in":[30,165],"real-world":[31],"settings":[32],"remains":[33],"uncertain,":[34],"interpretability":[37],"is":[38],"often":[39,142],"unexplored.":[40],"In":[41],"this":[42,196],"work,":[43],"we":[44,128,214],"investigate":[45],"whether":[46],"contemporary":[47],"detectors":[48,141,201],"genuinely":[49],"identify":[50],"machine":[51,69,153],"authorship":[52],"or":[53],"merely":[54],"exploit":[55],"dataset-specific":[56,145],"artefacts.":[57],"We":[58,193],"propose":[59],"an":[60,100,216],"interpretable":[61],"framework":[63],"that":[64,116,130,140,172,195,202,220],"integrates":[65],"linguistic":[66,94],"feature":[67],"engineering,":[68],"learning,":[70],"explainable":[72],"AI":[73,167,200],"techniques.":[74],"When":[75],"evaluated":[76],"on":[77,92,144,176],"two":[78],"prominent":[79],"corpora,":[81],"namely":[82],"PAN":[83],"CLEF":[84],"2025":[85],"COLING":[87],"2025,":[88],"our":[89],"model":[90],"trained":[91],"30":[93],"features":[95,134,171,182],"achieves":[96],"leaderboard-competitive":[97],"performance,":[98],"attaining":[99],"F1":[101],"score":[102],"0.9734.":[104],"However,":[105],"systematic":[106],"cross-domain":[107],"cross-generator":[109],"evaluation":[110],"reveals":[111],"substantial":[112],"generalisation":[113],"failure:":[114],"classifiers":[115],"excel":[117],"in-domain":[118,177],"degrade":[119],"significantly":[120],"under":[121],"distribution":[122],"shift.":[123],"Using":[124],"SHAP-":[125],"based":[126],"explanations,":[127],"show":[129],"most":[132,174,183],"influential":[133],"differ":[135],"markedly":[136],"between":[137],"datasets,":[138],"indicating":[139],"rely":[143],"stylistic":[146],"cues":[147],"rather":[148],"than":[149],"stable":[150],"signals":[151],"authorship.":[154],"Further":[155],"investigation":[156],"with":[157],"in-depth":[158],"error":[159],"analysis":[160],"exposes":[161],"fundamental":[163],"tension":[164],"linguistic-feature-based":[166],"detection:":[169],"are":[173,179,203],"discriminative":[175],"data":[178],"also":[180],"susceptible":[184],"to":[185],"domain":[186],"shift,":[187],"formatting":[188],"variation,":[189],"text-length":[191],"effects.":[192],"believe":[194],"knowledge":[197],"helps":[198],"build":[199],"robust":[204],"across":[205],"different":[206],"settings.":[207],"To":[208],"support":[209],"replication":[210],"practical":[212],"use,":[213],"release":[215],"open-source":[217],"Python":[218],"package":[219],"returns":[221],"both":[222],"predictions":[223],"instance-level":[225],"explanations":[226],"for":[227],"individual":[228],"texts.":[229]},"counts_by_year":[],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2026-03-26T00:00:00"}