Decoding Synthetic News: An Interpretable Multimodal Framework for the Classificationof News Articles in a Novel News Corpus

Decoding Synthetic News: An Interpretable Multimodal Framework for the Classificationof News Articles in a Novel News Corpus | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Decoding Synthetic News: An Interpretable Multimodal Framework for the Classificationof News Articles in a Novel News Corpus Michael Schlee, Gillian Kant, Christoph Ehrling, Benjamin Säfken, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4921351/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 07 Jul, 2025 Read the published version in Artificial Intelligence Review → Version 1 posted 9 You are reading this latest preprint version Abstract Recent advancements in Artificial Intelligence (AI), notably the development of Large Language Models (LLMs)and text-to-image diffusion models, have facilitated the creation of realistic textual content and images. Specifically, platformslike ChatGPT and Midjourney have simplified creation of high-quality text and visuals with minimal expertise and cost. Theincreasing sophistication of Generative AI presents challenges in ensuring the integrity of news, media, and information quality,making it increasingly difficult to distinguish between real and artificially generated textual and visual content. Our workaddresses this problem in two ways. First, by means of ChatGPT and Midjourney we create a comprehensive novel multimodalnews corpus named SyN24News based on the N24News corpus, on which we evaluate our model. Second, we develop a novelexplainable synthetic news detector for discriminating between real and synthetic news articles. We leverage a Neural AdditiveModel (NAM) like network structure that ensures effect separation by handling input data in separate subnetworks. Complexstructures and patterns are extracted by deep features from unstructured data, i.e. images and texts, using fine-tuned VGGand DistilBERT subnetworks. We ensure further explainability by individually processing carefully chosen handcrafted textand image features in simple Multilayer Perceptrons (MLPs), allowing for graphical interpretation of corresponding structuredeffects. Our findings indicate that textual information are the main drivers in the decision finding process. Structured textualeffects, particularly Flesch-Kincaid reading ease and sentiment, have a much higher influence on the classification outcomethan visual features such as dissimilarity and homogeneity. Generative AI Multimodal Synthetic Data Synthetic News Articles Explainable AI Neural Additive Models Feature Importance Analysis Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Published Journal Publication published 07 Jul, 2025 Read the published version in Artificial Intelligence Review → Version 1 posted Editorial decision: Revision requested 25 Oct, 2024 Reviews received at journal 25 Oct, 2024 Reviews received at journal 15 Oct, 2024 Reviewers agreed at journal 11 Oct, 2024 Reviewers agreed at journal 01 Oct, 2024 Reviewers invited by journal 01 Oct, 2024 Editor assigned by journal 26 Aug, 2024 Submission checks completed at journal 19 Aug, 2024 First submitted to journal 15 Aug, 2024 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4921351","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":354306096,"identity":"6d03df9c-8c6d-40b4-ae68-dce16e551974","order_by":0,"name":"Michael Schlee","email":"data:image/png;base64,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","orcid":"","institution":"University of Göttingen","correspondingAuthor":true,"prefix":"","firstName":"Michael","middleName":"","lastName":"Schlee","suffix":""},{"id":354306097,"identity":"80d9a70a-4147-4845-abde-a7b8e38dcae8","order_by":1,"name":"Gillian Kant","email":"","orcid":"","institution":"University of Göttingen","correspondingAuthor":false,"prefix":"","firstName":"Gillian","middleName":"","lastName":"Kant","suffix":""},{"id":354306098,"identity":"d6ade896-3531-417c-98dd-8eec293e5631","order_by":2,"name":"Christoph Ehrling","email":"","orcid":"","institution":"Finanz-DATA GmbH","correspondingAuthor":false,"prefix":"","firstName":"Christoph","middleName":"","lastName":"Ehrling","suffix":""},{"id":354306100,"identity":"e32bc3aa-450a-46ee-a58e-44f228577611","order_by":3,"name":"Benjamin Säfken","email":"","orcid":"","institution":"Clausthal University of Technology","correspondingAuthor":false,"prefix":"","firstName":"Benjamin","middleName":"","lastName":"Säfken","suffix":""},{"id":354306102,"identity":"624bad0b-8b28-4699-b466-9735a1c5b51a","order_by":4,"name":"Thomas Kneib","email":"","orcid":"","institution":"University of Göttingen","correspondingAuthor":false,"prefix":"","firstName":"Thomas","middleName":"","lastName":"Kneib","suffix":""}],"badges":[],"createdAt":"2024-08-15 21:19:42","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4921351/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4921351/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s10462-025-11188-9","type":"published","date":"2025-07-07T15:57:57+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":86699435,"identity":"c9181096-5800-4133-8832-d9e7ff40b19d","added_by":"auto","created_at":"2025-07-14 16:09:38","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2189847,"visible":true,"origin":"","legend":"","description":"","filename":"DecodingSyntheticNewsAnInterpretableMultimodalFrameworkfortheClassificationofNewsArticlesinaNovelNewsCorpus.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4921351/v1_covered_f5907d13-e8ae-4d94-aa3f-ad2c1806ce14.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Decoding Synthetic News: An Interpretable Multimodal Framework for the Classificationof News Articles in a Novel News Corpus","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"artificial-intelligence-review","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"aire","sideBox":"Learn more about [Artificial Intelligence Review](http://link.springer.com/journal/10462)","snPcode":"10462","submissionUrl":"https://submission.nature.com/new-submission/10462/3","title":"Artificial Intelligence Review","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Generative AI, Multimodal Synthetic Data, Synthetic News Articles, Explainable AI, Neural Additive Models, Feature Importance Analysis","lastPublishedDoi":"10.21203/rs.3.rs-4921351/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4921351/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Recent advancements in Artificial Intelligence (AI), notably the development of Large Language Models (LLMs)and text-to-image diffusion models, have facilitated the creation of realistic textual content and images. Specifically, platformslike ChatGPT and Midjourney have simplified creation of high-quality text and visuals with minimal expertise and cost. Theincreasing sophistication of Generative AI presents challenges in ensuring the integrity of news, media, and information quality,making it increasingly difficult to distinguish between real and artificially generated textual and visual content. Our workaddresses this problem in two ways. First, by means of ChatGPT and Midjourney we create a comprehensive novel multimodalnews corpus named SyN24News based on the N24News corpus, on which we evaluate our model. Second, we develop a novelexplainable synthetic news detector for discriminating between real and synthetic news articles. We leverage a Neural AdditiveModel (NAM) like network structure that ensures effect separation by handling input data in separate subnetworks. Complexstructures and patterns are extracted by deep features from unstructured data, i.e. images and texts, using fine-tuned VGGand DistilBERT subnetworks. We ensure further explainability by individually processing carefully chosen handcrafted textand image features in simple Multilayer Perceptrons (MLPs), allowing for graphical interpretation of corresponding structuredeffects. Our findings indicate that textual information are the main drivers in the decision finding process. Structured textualeffects, particularly Flesch-Kincaid reading ease and sentiment, have a much higher influence on the classification outcomethan visual features such as dissimilarity and homogeneity.","manuscriptTitle":"Decoding Synthetic News: An Interpretable Multimodal Framework for the Classificationof News Articles in a Novel News Corpus","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-09-16 07:23:03","doi":"10.21203/rs.3.rs-4921351/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2024-10-25T17:13:33+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-10-25T10:45:23+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-10-15T19:27:11+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"338520455687914211675646108329287924187","date":"2024-10-11T14:47:42+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"218971135336072373062588189918962192271","date":"2024-10-01T17:23:10+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2024-10-01T17:18:33+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-08-26T04:44:20+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-08-19T04:34:14+00:00","index":"","fulltext":""},{"type":"submitted","content":"Artificial Intelligence Review","date":"2024-08-15T21:18:07+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"artificial-intelligence-review","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"aire","sideBox":"Learn more about [Artificial Intelligence Review](http://link.springer.com/journal/10462)","snPcode":"10462","submissionUrl":"https://submission.nature.com/new-submission/10462/3","title":"Artificial Intelligence Review","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"ab67a1a7-e77e-4b56-9d76-aecd48520023","owner":[],"postedDate":"September 16th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-07-14T16:03:53+00:00","versionOfRecord":{"articleIdentity":"rs-4921351","link":"https://doi.org/10.1007/s10462-025-11188-9","journal":{"identity":"artificial-intelligence-review","isVorOnly":false,"title":"Artificial Intelligence Review"},"publishedOn":"2025-07-07 15:57:57","publishedOnDateReadable":"July 7th, 2025"},"versionCreatedAt":"2024-09-16 07:23:03","video":"","vorDoi":"10.1007/s10462-025-11188-9","vorDoiUrl":"https://doi.org/10.1007/s10462-025-11188-9","workflowStages":[]},"version":"v1","identity":"rs-4921351","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4921351","identity":"rs-4921351","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}