[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-84370-en":3,"doc-seo-84370-105":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},84370,13056703020460,"Valentina","https://ap-avatar.wpscdn.com/avatar/be000253dac470eee5d?_k=1778207105932848923",8,"Research & Report","The Context Access Divide Interaction-Level Architecture as a Complementary Dimension of Agentic Inequality","Sharp et al. (2025) frame agentic inequality through availability, quality, and quantity of AI agents. This work identifies a structurally important additional divide at the interaction level: whether an AI can autonomously retrieve context from a user’s personal or organizational knowledge corpus, or instead requires manual context attachment per query. The Context Access Divide (CAD) creates a qualitative threshold for knowledge work, shifting cognitive burdens back to humans and amplifying stratification. A probabilistic model and analysis of MCP and RAG formalize how manual curation collapses task success probability as corpora grow.","arXiv :2607 .08495v 1 [ cs .CY] 9 Jul 2026  \nThe Context Access Divide: Interaction-Level Architecture as a Complementary Dimension of Agentic Inequality  \nMasahiro Fujita  \nFaculty of Sociology, Kansai University  \n[m.fujita@kansai-u.ac.jp](m.fujita@kansai-u.ac.jp)  \nJuly 2026  \nPreprint—submitted to arXiv cs.CY  \nAbstract  \nSharp et al. [2025] introduce “agentic inequality” as a framework for analyzing disparities in access to AI agents across three dimensions: availability, quality, and quantity. These dimensions operate at the level of persons and organizations, characterizing who can access agents and at what capability. This paper identifies a structurally important divide that operates at a different level of analysis—the individual interaction—and that the Sharp et al. [2025] framework does not address: the difference between AI systems that can dynamically retrieve context from a user’s personal or organizational knowledge corpus and those that require manual context attachment by the user at each query. We term this the Context Access Divide (CAD) . For knowledge-intensive white-collar workers whose effective intellectual capital is distributed across tens of thousands of files, the CAD is not merely a convenience gap—it constitutes a qualitative threshold in AI usefulness. Below this threshold, the cognitive burden of context curation fallson the human, reproducing the very inefficiencies AI is supposed to eliminate. Because this interactionlevel architectural difference systematically advantages certain classes of workers and organizations, its distributional consequences aggregate to the societal level, making it a macro-consequential microvariable that complements Sharp et al.’s person-level analysis. We propose contextuality—the degree to which an AI system autonomously accesses a user’s accumulated knowledge capital—as a dimension of AI-mediated inequality that operates alongside, but at a different analytical level from, the availability, quality, and quantity dimensions. We formalize this divide by proposing a probabilistic model of human recall limits applied to context curation, grounded in the fan effect literature in cognitive psychology, and demonstrate formally how manual context attachment leads to a combinatorial collapse in task-success probability as corpus size and task conjunctivity grow—a collapse from which dynamic context retrieval architectures are structurally insulated. We examine the technical basis of this divide in the Model Context Protocol (MCP) and retrieval-augmented generation (RAG) architectures, and analyze its implications for the stratification of knowledge work.  \nKeywords: agentic AI, AI inequality, knowledge work, Model Context Protocol, retrieval-augmented generation, digital divide, white-collar labor  \n1 Introduction  \nThe emergence of autonomous AI agents—systems capable of planning, tool use, and multi-step task execution—marks a qualitative shift beyond conversational AI chatbots. As Sharp et al. [2025] observe, this shift is not merely technological but distributional: who can access capable agents, and at what level of capability, will shape the distribution of power and opportunity in economic and social life.  \nThe Sharp et al. [2025] framework identifies three core dimensions along which agentic inequality manifests: availability (whether one can access any agent at all), quality (the capability level of the agent one  \ncan access), and quantity (the number of agents one can deploy simultaneously) . This framework represents an important conceptual advance over earlier digital divide scholarship, which focused on binary access to infrastructure.  \nYet the framework, as currently formulated, operates at a level of analysis that may obscure a structurally significant divide operating within the “availability” and “quality” dimensions. Two users may both have nominal access to the same AI platform—the same subscription tier, the same underlying model—and yet experience","cbCaikIjQpABtrCR","https://ap.wps.com/l/cbCaikIjQpABtrCR","pdf",283874,1,19,"English","en",105,"# Abstract\n# Introduction","[{\"question\":\"What does the paper add to Sharp et al.'s framework of agentic inequality?\",\"answer\":\"It adds the Context Access Divide (CAD), an interaction-level factor describing whether an AI system can autonomously retrieve relevant context from a user’s corpus or requires manual context attachment per query.\"},{\"question\":\"Why is the Context Access Divide especially consequential for knowledge-intensive white-collar work?\",\"answer\":\"For workers whose intellectual capital is spread across tens of thousands of files, dynamic context retrieval becomes a qualitative threshold for AI usefulness, while manual curation imposes cognitive burden that undermines the goal of reducing inefficiencies.\"},{\"question\":\"How does the paper explain the effect of manual context attachment on task success?\",\"answer\":\"It formalizes the divide with a probabilistic model grounded in the fan effect, showing that manual context attachment leads to a combinatorial collapse in task-success probability as corpus size and task conjunctivity increase.\"}]",1784195153,48,{"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},"the-context-access-divide-interaction-level-architecture-as-a-complementary-dimension-of-agentic-inequality","",{"@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/the-context-access-divide-interaction-level-architecture-as-a-complementary-dimension-of-agentic-inequality/84370/",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 does the paper add to Sharp et al.'s framework of agentic inequality?","Question",{"text":75,"@type":76},"It adds the Context Access Divide (CAD), an interaction-level factor describing whether an AI system can autonomously retrieve relevant context from a user’s corpus or requires manual context attachment per query.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"Why is the Context Access Divide especially consequential for knowledge-intensive white-collar work?",{"text":80,"@type":76},"For workers whose intellectual capital is spread across tens of thousands of files, dynamic context retrieval becomes a qualitative threshold for AI usefulness, while manual curation imposes cognitive burden that undermines the goal of reducing inefficiencies.",{"name":82,"@type":73,"acceptedAnswer":83},"How does the paper explain the effect of manual context attachment on task success?",{"text":84,"@type":76},"It formalizes the divide with a probabilistic model grounded in the fan effect, showing that manual context attachment leads to a combinatorial collapse in task-success probability as corpus size and task conjunctivity increase.","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},[93,97,101,105,110,115,120,123,128,131,135],{"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":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":45,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":45,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":45,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":45,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":21,"doc_module":4,"doc_module_name":45,"category_name":136,"show_sort_weight":106,"slug":137},"General","general"]