[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-646895-105":59,"doc-detail-646895-en":129},{"code":4,"msg":5,"data":6},0,"success",[7,13,18,23,28,33,38,43,48,51,55],{"id":8,"doc_module":4,"doc_module_name":9,"category_name":10,"show_sort_weight":11,"slug":12},1,"Document","Story & Novel",90,"story-novel",{"id":14,"doc_module":4,"doc_module_name":9,"category_name":15,"show_sort_weight":16,"slug":17},2,"Literature",80,"literature",{"id":19,"doc_module":4,"doc_module_name":9,"category_name":20,"show_sort_weight":21,"slug":22},4,"Exam",70,"exam",{"id":24,"doc_module":4,"doc_module_name":9,"category_name":25,"show_sort_weight":26,"slug":27},5,"Comic",60,"comic",{"id":29,"doc_module":4,"doc_module_name":9,"category_name":30,"show_sort_weight":31,"slug":32},6,"Technology",50,"technology",{"id":34,"doc_module":4,"doc_module_name":9,"category_name":35,"show_sort_weight":36,"slug":37},7,"Healthcare",40,"healthcare",{"id":39,"doc_module":4,"doc_module_name":9,"category_name":40,"show_sort_weight":41,"slug":42},8,"Research & Report",30,"research-report",{"id":44,"doc_module":4,"doc_module_name":9,"category_name":45,"show_sort_weight":46,"slug":47},9,"Religion & Spirituality",20,"religion-spirituality",{"id":46,"doc_module":4,"doc_module_name":9,"category_name":49,"show_sort_weight":46,"slug":50},"World Cup","world-cup",{"id":52,"doc_module":4,"doc_module_name":9,"category_name":53,"show_sort_weight":52,"slug":54},10,"Lifestyle","lifestyle",{"id":56,"doc_module":4,"doc_module_name":9,"category_name":57,"show_sort_weight":24,"slug":58},19,"General","general",{"code":4,"msg":60,"data":61},"ok",{"site_id":62,"language":63,"slug":64,"title":65,"keywords":66,"description":67,"schema_data":68,"social_meta":122,"head_meta":124,"extra_data":126,"updated_unix":128},105,"en","automated-profile-inference-with-language-model-agents-paper","Automated Profile Inference with Language Model Agents - Paper","","Automated problem-solving by large language model (LLM) agents brings new privacy threats, specifically automated profile inference. The work studies how an adversary can instruct LLMs to collect and extract sensitive personal attributes from publicly available activity on pseudonymous platforms. It proposes AutoProfiler, a four-agent collaborative framework for retrieving activities, extracting attributes, and resolving inconsistencies. Experiments on two real datasets and one synthetic dataset show high effectiveness and efficiency, with inferred attributes that are identifiable and sensitive, creating serious privacy risks. Mitigation and awareness strategies are discussed.",{"@graph":69,"@context":121},[70,84,104],{"@type":71,"itemListElement":72},"BreadcrumbList",[73,77,79,82],{"item":74,"name":75,"@type":76,"position":8},"https://docshare.wps.com","Home","ListItem",{"item":78,"name":9,"@type":76,"position":14},"https://docshare.wps.com/document/",{"item":80,"name":40,"@type":76,"position":81},"https://docshare.wps.com/document/research-report/",3,{"item":83,"name":65,"@type":76,"position":19},"https://docshare.wps.com/document/automated-profile-inference-with-language-model-agents-paper/646895/",{"url":83,"name":65,"@type":85,"image":86,"author":91,"headline":65,"publisher":94,"fileFormat":97,"inLanguage":63,"description":67,"dateModified":98,"datePublished":98,"encodingFormat":97,"isAccessibleForFree":99,"interactionStatistic":100},"DigitalDocument",{"url":87,"@type":88,"width":89,"height":90},"https://docshare.wps.com/thumbnails/automated-profile-inference-with-language-model-agents-paper/646895.png","ImageObject",300,407,{"name":92,"@type":93},"jayni","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-10-23",true,{"@type":101,"interactionType":102,"userInteractionCount":4},"InteractionCounter",{"@type":103},"ViewAction",{"@type":105,"mainEntity":106},"FAQPage",[107,113,117],{"name":108,"@type":109,"acceptedAnswer":110},"What is automated profile inference in this paper?","Question",{"text":111,"@type":112},"It is an attack where an adversary instructs LLMs to automatically collect and extract sensitive personal attributes from publicly visible user activities on pseudonymous platforms, enabling detailed user profiles and privacy breaches.","Answer",{"name":114,"@type":109,"acceptedAnswer":115},"How does AutoProfiler work?",{"text":116,"@type":112},"AutoProfiler is a multi-agent system with four specialized LLM agents: Strategist, Retriever, Extractor, and Summarizer. They collaborate to retrieve online activities, extract personal details, and refine results to improve reliability.",{"name":118,"@type":109,"acceptedAnswer":119},"What do the experiments show about the risk posed by AutoProfiler?",{"text":120,"@type":112},"On two real-world datasets and one synthetic dataset, AutoProfiler is effective and efficient, producing inferred attributes that are both identifiable and sensitive. This increases practical privacy risk, including potential de-anonymization.","https://schema.org",{"og:url":83,"og:type":123,"og:title":65,"og:site_name":95,"og:description":67},"article",{"robots":125,"canonical":83},"index,follow",{"doc_id":127,"site_id":62},646895,1792773527,{"code":4,"msg":5,"data":130},{"doc_id":127,"user_id":131,"nickname":92,"user_avatar":132,"doc_module":4,"category_id":39,"category_name":40,"doc_title":65,"doc_description":67,"doc_content":133,"file_id":134,"file_url":135,"file_type":136,"file_size":137,"view_count":4,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":138,"language":139,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":140,"faqs":141,"seo_title":142,"seo_description":67,"update_tm":128,"read_time":26},3985747859343,"https://ap-avatar.wpscdn.com/davatar_276721f389ce27ea32af1340a28f341c","Automated Profile Inference with Language Model Agents  \nYuntao Du1 , Zitao Li2 , Bolin Ding2 , Yaliang Li2 , Hanshen Xiao1,3 , Jingren Zhou2 , Ninghui Li1  \n1Purdue University, 2Alibaba Group, 3Nvidia  \nCorrespondence: [ytdu@purdue.edu](ytdu@purdue.edu)  \narXiv :2505 . 12402v2 [ cs .CR] 14 Apr 2026  \nAbstract  \nImpressive progress has been made in automated problem-solving by the collaboration of large language model (LLM) based agents.  \nHowever, these automated capabilities also open avenues for malicious applications. In this paper, we study a new threat that LLMs pose to online pseudonymity, called automated profile inference, where an adversary can instruct LLMs to automatically collect and extract sensitive personal attributes from publicly available user activities on pseudonymous platforms. We also introduce an automated profiling framework called AutoProfiler to demonstrate and assess the feasibility of such attacks in real-world scenarios. AutoProfiler consists of four specialized LLM agents that work collaboratively to retrieve and process user online activities and generate a profile with extracted personal information. Experimental results on two real-world datasets and one synthetic dataset show that AutoProfiler is highly effective and efficient, and the inferred attributes are both identifiable and sensitive, posing significant privacy risks.  \nWe explore mitigation strategies from different perspectives and advocate for increased public awareness of this emerging privacy threat.  \n1 Introduction  \nIn recent years, large language models (LLMs) have become increasingly capable, enabling autonomous agents that can enhance and replicate complex human workflows and demonstrating strong performance across diverse tasks (Yang et al., 2024b ; Park et al., 2023 ; Boiko et al., 2023) . However, these same capabilities have raised concerns due to the potential for malicious applications (Fang et al., 2024 ; Law, 2023) . Notably, there has been a rise in privacy concerns regarding LLMs. In addition to widely studied data privacy risks (Carlini et al., 2021 ; Lukas et al., 2023), LLMs can violate individuals’ privacy in unexpected ways. For instance, studies (Mireshghallah  \nFigure 1: Illustration of automated profile inference: an adversary instructs AutoProfiler to autonomously collect and analyze users’ online activities, infer personal attributes, and generate detailed user profiles that may result in privacy breaches.  \net al., 2024 ; Staab et al., 2024) show that an adversary can use LLMs to steal/infer users’ information by steering chat conversations.  \nIn this paper, we study an emerging privacy threat that LLMs pose to online pseudonymity. As shown in Figure 1, an adversary can, with the help of LLMs, automatically extract sensitive personal information from the publicly visible online activities (e.g., posts and comments) of a user on a pseudonymous platform (e.g., Reddit) . When a user has conducted substantial online activities, the adversary can even build a detailed description of the user. We call this attack automated profile inference. Relying only on a user’s public online activities, we find that the resulting privacyinfringing inferences can reveal highly private information about the user, and the inferred profiles can be exploited to facilitate privacy breaches, such as de-anonymization. While similar attacks can be carried out using traditional profiling approaches (Douglas et al., 1986 ; Estival et al., 2007), doing so requires significant manual effort and expertise. Automated profile inference puts the same capability in the hands of any adversary, changing the magnitude of such privacy threats.  \nSince users’ online activities on the pseudonymous platform are usually ambiguous, inconsistent, and full of superficially insensitive information, we find that simply feeding these texts to an LLM and instructing it to generate a profile  \nstruggles to extract implicit personal details. To address thi","cbCaib2kX8HnH6M1","https://ap.wps.com/l/cbCaib2kX8HnH6M1","pdf",1852199,24,"English","# Abstract\n# 1 Introduction\n## Automated profile inference threat\n## AutoProfiler multi-agent framework\n## Evaluation on real and synthetic datasets\n## Mitigation strategies","[{\"question\":\"What is automated profile inference in this paper?\",\"answer\":\"It is an attack where an adversary instructs LLMs to automatically collect and extract sensitive personal attributes from publicly visible user activities on pseudonymous platforms, enabling detailed user profiles and privacy breaches.\"},{\"question\":\"How does AutoProfiler work?\",\"answer\":\"AutoProfiler is a multi-agent system with four specialized LLM agents: Strategist, Retriever, Extractor, and Summarizer. They collaborate to retrieve online activities, extract personal details, and refine results to improve reliability.\"},{\"question\":\"What do the experiments show about the risk posed by AutoProfiler?\",\"answer\":\"On two real-world datasets and one synthetic dataset, AutoProfiler is effective and efficient, producing inferred attributes that are both identifiable and sensitive. This increases practical privacy risk, including potential de-anonymization.\"}]","Automated Profile Inference with Language Model Agents - Paper | PDF"]