{"resourceId":"ieee-spectrum-ai-faked-paper-ieee-access-integrity-checks-20261027","versions":[{"version":"external-8d8fbd0be64a3af294d9d5005951e8994012e2ccff67dd540e2106201868045f","resource":{"id":"ieee-spectrum-ai-faked-paper-ieee-access-integrity-checks-20261027","title":"IEEE Spectrum war-games an AI-faked paper and an autonomous-system paper against IEEE Access: integrity checks caught both, on references","organization":"IEEE Spectrum (Benjamin Skuse) with IEEE Access managing editor Jenny Mahoney and IEEE research-integrity staff","sector":"Scholarly publishing and research integrity","geography":"International (IEEE)","publishedAt":"27 October 2026 (IEEE Spectrum feature; listed in Retraction Watch weekend reads, 31 Oct 2026)","publicationDate":"2026-10-27","eventDate":null,"sourceName":"IEEE Spectrum","sourceLabel":"Benjamin Skuse, \"I Used AI to Fake a Scientific Paper. Here's What Happened,\" IEEE Spectrum, 27 Oct 2026","sourceUrl":"https://spectrum.ieee.org/ai-generated-scientific-papers","evidenceClass":"independent-reporting","outcomeClass":"cautionary","topics":["knowledge-work","governance-procurement","developers-agents"],"finding":"A journalist with no special AI expertise used free ChatGPT and Gemini, across about two dozen prompts, to produce a realistic-looking optical-solitons paper describing research never done, then, with IEEE's agreement, submitted it to IEEE Access. Automated integrity checks flagged fabricated and unverifiable references; STM Integrity Hub found 12 and a MolecularConnections checker could not validate 24. A second paper produced by the autonomous research system FARS had real references but was flagged for heavy reliance on non-peer-reviewed preprints. An external expert found the first paper professionally written but flagged debatable physics claims. IEEE Access receives about 60,000 submissions a year and accepts about 20%; reviewer invitations per article are up 50% since 2018. IEEE's publishing-ethics director said today's detection tools may not work on tomorrow's AI-created papers.","sledRelevance":"Concrete evidence for research offices and libraries that current publisher defenses catch AI fabrication mainly through reference checks, and that autonomous research systems already avoid that tell. Newly relevant now as it circulated in Retraction Watch's 31 October roundup alongside the IEEE bulk retraction.","evidence":"Read the article through the FARS section. A two-paper journalistic test, not a study; journal staff knew of the experiment. Opening statistics come from linked studies that were not opened.","architectureImplications":"Publisher screening relies on reference-validation and integrity tools; institutions could run similar checks before submission.","governanceImplications":"Reference validation alone will not catch AI-conducted research with real citations; disclosure rules and expert review remain the backstop.","securityPrivacyImplications":"Not addressed.","caveats":"Published 27 Oct, before the previous run, and included because it circulated on 31 Oct. IEEE Spectrum is IEEE's own magazine reporting on an IEEE journal, a conflict of interest. Two papers, cooperative journal, one field.","streamIds":["research"],"roles":{"sales":"Lead with pre-submission integrity checks as research AI use grows. IEEE Spectrum reported on 27 October that an AI-faked paper was caught at IEEE Access because integrity tools found 12 to 24 bad references, while a paper from an autonomous research system had real references and was flagged only for heavy preprint use. Research integrity officers need checks beyond citations. Offer a bounded review of pre-submission screening for one college. Do not claim detection tools reliably catch AI fabrication.","engineering":"Frame fit around reference validation plus provenance. The journal's tools caught fabricated citations, but the autonomous-system paper passed that test. Prerequisites are access to reference-validation services, institutional repositories that hold code and data, and disclosure forms that capture AI use. Validate by running known AI-generated and human papers through the chosen checks and measuring which flaws each catches, especially papers with real references and fabricated results.","delivery":"Treat delivery as an integrity workflow, not a detector purchase. Work covers disclosure guidance, a pre-submission checklist, reference and data provenance checks, and reviewer training. Ownership sits with the research integrity office, with libraries running checks. Governance checkpoints cover disclosure review and escalation of suspected fabrication. Acceptance criteria are proposals, such as all flagged references resolved before submission and a tested escalation path, not results IEEE reported."},"roleDetails":{"sales":{"customerProblem":"AI makes convincing fake papers easy, and journals catch them mainly through bad references that better tools avoid.","stakeholders":["Research integrity officer","University librarian","Vice president for research","Department chairs"],"discoveryQuestions":["Do researchers disclose AI use before submission?","Does the library run reference or integrity checks?","How are suspected fabrications escalated?"],"valueHypothesis":"Pre-submission reference and provenance checks may catch AI-related errors before they reach journals, but not all fabrication.","potentialEngagement":"A review of pre-submission integrity screening for one college.","unsupportedClaims":["Detectors reliably catch AI papers","Results generalize beyond two papers","Peer review can absorb the load"]},"engineering":{"fit":"Fits universities with library publishing support and a research integrity office.","architectureIntegration":"Reference-validation services plus repository links to code and data, tied to AI-use disclosure forms.","prerequisites":["Reference-validation tool access","Code and data repository","AI-use disclosure form"],"constraints":["Tools miss real-reference fabrication","Two-case evidence","Publisher conflict of interest"],"security":"Keep unpublished manuscripts in approved tools only, not public AI services.","proofOfValue":["Test checks on known AI and human papers","Measure catch rates by flaw type","Pilot with one department"]},"delivery":{"work":"Write disclosure guidance and a checklist, run reference and provenance checks, and train reviewers.","dependencies":["Library services","Integrity tool licenses","Faculty participation"],"ownership":"Research integrity office with the university library.","skillsAdoption":"Researchers learn to verify AI-assisted references; staff learn to interpret tool flags.","governanceCheckpoints":["Disclosure reviewed before submission","Escalation path for suspected fabrication","Annual review of tools"],"acceptanceCriteria":["All flagged references resolved before submission (proposal)","Escalation path tested","Pilot department trained"],"risks":["False assurance from tools","Researcher resistance","Confidential manuscripts in public AI tools"]}},"retrievedAt":"2026-11-02T08:38:58-06:00","enrichedAt":"2026-11-02T08:38:59-06:00","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Volunteer reviewers face growing load; the article cites a 50% rise in invitations per article since 2018.","procurementImplications":"Names commercial integrity tools (STM Integrity Hub, MolecularConnections) used by the journal.","operatingModelImplications":"Suggests research offices pair AI-use disclosure with pre-submission reference and provenance checks.","peerReview":"unknown","sourceVerification":{"openedUrl":"https://spectrum.ieee.org/ai-generated-scientific-papers","referenceExcerpt":"As it turned out, I was filled with relief when the paper got picked up quickly for having hallucinated references, otherwise known as ghost citations.","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}