[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-85306-en":3,"doc-seo-85306-105":30,"detail-sidebar-cat-0-en-105":92},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":21,"is_downloadable":21,"audit_status":21,"page_count":22,"language":23,"language_code":24,"site_id":25,"html_lang":24,"table_of_contents":26,"faqs":27,"seo_title":13,"seo_description":14,"update_tm":28,"read_time":29},85306,687197207919,"Theodora","https://ap-avatar.wpscdn.com/avatar/a000253d6f5f7c60be?x-image-process=image/resize,m_fixed,w_180,h_180&k=1779446848396160552",8,"Research & Report","Automated Textbook Auditing with Multi-Agent LLM Systems","Ensuring the quality of educational materials requires auditing textbooks for factual accuracy, domain-specific technical correctness, and linguistic quality simultaneously—capabilities that general grammar checkers cannot cover. AI Textbook Auditor is a modular multi-agent pipeline that ingests a textbook PDF and outputs a structured, human-reviewable report. Two tracks detect factual/technical problems and grammar issues, with a Judge Agent filtering false positives. The system supports vision-native rendering and PDF text extraction, uses domain-adaptable prompts, and is validated on Romanian CS and humanities textbooks.","Automated Textbook Auditing with Multi-Agent LLM Systems  \nCiprian Cristescu1 , Adrian-Marius Dumitran 1 , Angela-Liliana Dumitran2 and Gabriel S, tefan 1  \n1 University of Bucharest, Bucharest, Romania  \n2 Dimitrie Cantemir Christian University, Bucharest, Romania  \nAbstract  \nEnsuring the quality of educational materials requires more than standard proofreading: textbooks must be audited for factual accuracy, domain-specific technical correctness, and linguistic quality simultaneously — atask that general-purpose grammar checkers cannot address. We present AI Textbook Auditor, a modular multi-agent pipeline for automated quality assurance of educational materials across subject domains. The system accepts a textbook PDF and produces a structured, human-reviewable report via two analysis tracks: a Factual and Technical Track in which an ensemble of specialized LLM agents detects factual inaccuracies, code errors, incorrect definitions, and conceptual inconsistencies, augmented with web search for humanities domains; and a Grammar Track operating PDF-natively to preserve diacritical encoding. A Judge Agent filters false positives using domain-specific rules before presenting findings to a human reviewer. The pipeline supports two ingestion modes —vision-native page rendering and PyMuPDF text extraction — and is domain-adaptable via custom prompts encoding subject-specific error taxonomies. We demonstrate the system on two Romanian upper-secondary textbooks: a CS textbook (56 technical findings across seven categories, with an expert-validated precision of 62.5%) and a history and social sciences textbook (72 findings spanning factual errors, ideological bias, and grammar) . The system is designed as a triage tool that reduces the manual effort of locating candidate issues, with human expert validation required before any editorial action.  \n1. Introduction  \nHigh-quality textbooks are foundational to effective teaching, yet ensuring their quality at scale remains an open challenge. General-purpose grammar checkers address only the linguistic dimension and are blind to the error types that matter most in technical and scientific textbooks: a misplaced stream operator in a C/C++ example, an extra mathematical symbol, a historical date that contradicts the established record, or a mountain’s altitude stated incorrectly in a geography chapter.  \nWe present AI Textbook Auditor, a modular multi-agent system for automated quality assurance of educational materials. An ensemble of specialized LLM Agents analyzes textbook content for factual inaccuracies and domain-specific technical errors, followed by a Judge Agent that filters false positives before presenting findings to a human reviewer. The system is domain-adaptable with the help of a domain expert via a custom prompt specifying the subject, error taxonomy, and negative constraints; for humanities textbooks it is augmented with web search for claim verification, while for technical subjects it relies on parametric model knowledge. We demonstrate the system on Romanian upper-secondary textbooks across two subject domains.  \nThe main contributions of this work are: (i) a modular multi-agent auditing pipeline with domain adaptation via custom prompts; (ii) a Judge Agent for domain-aware false-positive filtering without additional human annotation in the detection loop; (iii) a unified architecture supporting web-search mode for humanities and parametric-knowledge mode for technical domains; (iv) preliminary evaluation on two Romanian upper-secondary textbooks spanning CS and humanities domains.  \niTextbooks’26: Seventh Workshop on Intelligent Textbooks, June 28, 2026, Seoul, Republic of Korea  \n$ [cristescuciprian.cc@gmail.com](cristescuciprian.cc@gmail.com) (C. Cristescu); [marius.dumitran@unibuc.ro](marius.dumitran@unibuc.ro) (A. Dumitran); [dumitranangela@gmail.com](dumitranangela@gmail.com)  \n(A. Dumitran); [gabrielstefan04@gmail.com](gabrielstefan04@gmail.com) (G. S, tefan)  \n","cbCaiqPLDurJdOrK","https://ap.wps.com/l/cbCaiqPLDurJdOrK","pdf",812540,6,1,7,"English","en",105,"# Introduction\n## Contributions\n# Related Work\n## Multi-agent LLM architectures\n## Code error detection and limitations\n## Web-augmented fact-checking\n# System Architecture","[{\"question\":\"What types of textbook quality issues does AI Textbook Auditor target?\",\"answer\":\"It targets factual inaccuracies, domain-specific technical errors, code errors and incorrect definitions, conceptual inconsistencies, and linguistic quality problems in the grammar track.\"},{\"question\":\"How does the multi-agent pipeline reduce false positives?\",\"answer\":\"A dedicated Judge Agent applies domain-specific rules to filter out false positives before presenting findings to a human reviewer.\"},{\"question\":\"How is domain verification handled differently for humanities vs technical subjects?\",\"answer\":\"Humanities claims are verified with optional web search, while technical subjects rely on parametric model knowledge based on stable language standards and algorithmic theory.\"}]",1784202371,18,{"code":4,"msg":31,"data":32},"ok",{"site_id":25,"language":24,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":87,"head_meta":89,"extra_data":91,"updated_unix":28},"automated-textbook-auditing-with-multi-agent-llm-systems","",{"@graph":36,"@context":86},[37,54,69],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/automated-textbook-auditing-with-multi-agent-llm-systems/85306/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":24,"description":14,"dateModified":62,"datePublished":63,"encodingFormat":61,"isAccessibleForFree":64,"interactionStatistic":65},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-07-23","2026-07-16",true,{"@type":66,"interactionType":67,"userInteractionCount":20},"InteractionCounter",{"@type":68},"ViewAction",{"@type":70,"mainEntity":71},"FAQPage",[72,78,82],{"name":73,"@type":74,"acceptedAnswer":75},"What types of textbook quality issues does AI Textbook Auditor target?","Question",{"text":76,"@type":77},"It targets factual inaccuracies, domain-specific technical errors, code errors and incorrect definitions, conceptual inconsistencies, and linguistic quality problems in the grammar track.","Answer",{"name":79,"@type":74,"acceptedAnswer":80},"How does the multi-agent pipeline reduce false positives?",{"text":81,"@type":77},"A dedicated Judge Agent applies domain-specific rules to filter out false positives before presenting findings to a human reviewer.",{"name":83,"@type":74,"acceptedAnswer":84},"How is domain verification handled differently for humanities vs technical subjects?",{"text":85,"@type":77},"Humanities claims are verified with optional web search, while technical subjects rely on parametric model knowledge based on stable language standards and algorithmic theory.","https://schema.org",{"og:url":52,"og:type":88,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":90,"canonical":52},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":93},[94,98,102,106,111,115,119,122,127,130,134],{"id":21,"doc_module":4,"doc_module_name":46,"category_name":95,"show_sort_weight":96,"slug":97},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":99,"show_sort_weight":100,"slug":101},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":103,"show_sort_weight":104,"slug":105},"Exam",70,"exam",{"id":107,"doc_module":4,"doc_module_name":46,"category_name":108,"show_sort_weight":109,"slug":110},5,"Comic",60,"comic",{"id":20,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},"Technology",50,"technology",{"id":22,"doc_module":4,"doc_module_name":46,"category_name":116,"show_sort_weight":117,"slug":118},"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":120,"slug":121},30,"research-report",{"id":123,"doc_module":4,"doc_module_name":46,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":46,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":107,"slug":137},19,"General","general"]