[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-84618-en":3,"doc-seo-84618-105":29,"detail-sidebar-cat-0-en-105":91},{"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":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},84618,1649267921044,"Ava Thompson","https://us-avatar.wpscdn.com/avatar/1800007509477c92dfb?_k=1782875107921204101",8,"Research & Report","A Practice Auditing Framework for Large Language Model Use","Large language models increasingly support knowledge acquisition, code generation, image creation, and academic writing, yet structured answers may outpace real domain practice. This paper proposes a practice auditing framework to explain how outputs can convert compressed collective experience into users’ pseudo-rational cognition. It analyzes cognitive loops and governance risks when unaudited AI content enters memory systems, retrieval spaces, future contexts, and detection workflows. It then prescribes auditing steps such as requirements definition, evidence-source auditing, practical validation, reverse questioning, logging, version control, rollback, and renewed cognition.","A Practice Auditing Framework for Large Language Model Use: Collective Empiricism, Pseudo-Rational Cognition, and Governance of AI-Generated Content  \nYang Zhao [zyxiaoqi7@gmail. com](zyxiaoqi7@gmail. com)  \nYingshuo Li[lys20020624@gmail. com](lys20020624@gmail. com)  \narXiv :2607 .0 1248v 1 [ cs .CY] 2 Jun 2026  \nZeyu Zhang∗  \nSchool of Information and Electronics, Beijing Institute of Technology  \n[zhangzeyu@bit. edu. cn](zhangzeyu@bit. edu. cn)  \nAbstract  \nLarge language models are increasingly used for knowledge acquisition, code generation, image creation, academic writing, and agent-based systems. In these settings, users can obtain highly structured answers, plans, and judgments without sufficient domain practice. However, structured output does not mean that the user has completed a rational understanding of the concrete problem. This paper addresses LLM interaction and governance of AIgenerated content by proposing a practice auditing framework. The framework explainshow AI outputs may be transformed from compressed collective experience into user-level pseudo-rational cognition, and how unaudited AI-generated content may create cognitive loops and governance risks when it enters future contexts, memory systems, retrieval spaces, or detection mechanisms.  \nThe paper first introduces the concept of collective empiricism to describe the mechanism by which large language models, without subject-centered practice, statistically learn, semantically reorganize, and contextually adapt large-scale human experience materials into outputs that appear empirical and rational. It then defines pseudo-rational cognition as a condition in which users mistake AI-generated structured expression for their own rational understanding despite lacking sufficient sensory materials, practice feedback, and conditional analysis. On this basis, the paper analyzes the illusion of AI subjectivity, subjectivity structures embedded in input materials, template loops in AI-AI conversations, statistical misjudgment in AIGC detection, and memory pollution or self-reinforcement caused by AI-generated content entering future retrieval spaces.  \nTo mitigate these risks, the paper proposes a practice auditing process for AI use: requirement definition, problem-boundary identification, evidence-source auditing, practical validation, reverse questioning, logging, version management, rollback mechanisms, and renewed cognition. The framework does not deny the productivity value of AI. Instead, it argues that LLM outputs should be returned to verifiable, reproducible, and intervenable processes of practice. The contribution of this paper is to provide an analyzable and auditable theoretical framework for cognitive risks in LLM use, AI-generated content governance, longterm memory systems, and human-AI interaction.  \nKeywords: Large Language Models; AI-Generated Content; Practice Auditing; Collective Empiricism; Pseudo-Rational Cognition; Memory Pollution; AI Governance; Human-AI Interaction  \n∗ Corresponding author  \n1 Introduction  \nLarge language models have increasingly become general-purpose interfaces for knowledge acquisition, code generation, image creation, academic writing, and agent-based automation. In these settings, users no longer interact with AI systems merely as search tools, but as systems that generate structured plans, explanations, judgments, code, summaries, and reusable intermediate artifacts. These outputs may further enter future prompts, memory modules, retrieval-augmented generation pipelines, AI-agent skill libraries, or AIGC detection workflows. Therefore, the key problem is no longer only whether an AI output is factually correct in a single turn, but how AI-generated structured content participates in the user’s longer process of understanding, decision-making, validation, and revision.  \nThis shift creates a governance problem. A user who does not understand the technical structure behind an artifact may still generate code, ima","cbCailWRDO5CxIiA","https://ap.wps.com/l/cbCailWRDO5CxIiA","pdf",255377,1,20,"English","en",105,"# Introduction\n## Contributions","[{\"question\":\"What problem does the paper address in large language model use?\",\"answer\":\"The paper focuses on governance and interaction risks: structured AI outputs may seem rational without users completing the necessary practice, validation, and conditional analysis across longer understanding and decision processes.\"},{\"question\":\"How do collective empiricism and pseudo-rational cognition relate to cognitive risk?\",\"answer\":\"Collective empiricism describes how LLMs transform large-scale human experience materials into outputs that appear empirical and rational. Pseudo-rational cognition occurs when users mistake these structured expressions for their own rational understanding despite lacking sufficient sensory evidence, practice feedback, and conditional analysis.\"},{\"question\":\"What does the proposed practice auditing framework include?\",\"answer\":\"It includes requirement definition, problem-boundary identification, evidence-source auditing, practical validation, reverse questioning, logging, version management, rollback mechanisms, and renewed cognition to ensure AI outputs remain verifiable, reproducible, and intervenable.\"}]",1784197165,50,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":27},"a-practice-auditing-framework-for-large-language-model-use","",{"@graph":35,"@context":85},[36,53,68],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/a-practice-auditing-framework-for-large-language-model-use/84618/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-17","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does the paper address in large language model use?","Question",{"text":75,"@type":76},"The paper focuses on governance and interaction risks: structured AI outputs may seem rational without users completing the necessary practice, validation, and conditional analysis across longer understanding and decision processes.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How do collective empiricism and pseudo-rational cognition relate to cognitive risk?",{"text":80,"@type":76},"Collective empiricism describes how LLMs transform large-scale human experience materials into outputs that appear empirical and rational. Pseudo-rational cognition occurs when users mistake these structured expressions for their own rational understanding despite lacking sufficient sensory evidence, practice feedback, and conditional analysis.",{"name":82,"@type":73,"acceptedAnswer":83},"What does the proposed practice auditing framework include?",{"text":84,"@type":76},"It includes requirement definition, problem-boundary identification, evidence-source auditing, practical validation, reverse questioning, logging, version management, rollback mechanisms, and renewed cognition to ensure AI outputs remain verifiable, reproducible, and intervenable.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,114,119,122,126,129,133],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":45,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":45,"category_name":112,"show_sort_weight":28,"slug":113},6,"Technology","technology",{"id":115,"doc_module":4,"doc_module_name":45,"category_name":116,"show_sort_weight":117,"slug":118},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":120,"slug":121},30,"research-report",{"id":123,"doc_module":4,"doc_module_name":45,"category_name":124,"show_sort_weight":21,"slug":125},9,"Religion & Spirituality","religion-spirituality",{"id":21,"doc_module":4,"doc_module_name":45,"category_name":127,"show_sort_weight":21,"slug":128},"World Cup","world-cup",{"id":130,"doc_module":4,"doc_module_name":45,"category_name":131,"show_sort_weight":130,"slug":132},10,"Lifestyle","lifestyle",{"id":134,"doc_module":4,"doc_module_name":45,"category_name":135,"show_sort_weight":106,"slug":136},19,"General","general"]