[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-83605-en":3,"doc-seo-83605-105":29,"detail-sidebar-cat-0-en-105":90},{"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":4,"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},83605,16904993612988,"Olivia Brown","https://ap-avatar.wpscdn.com/davatar_a8503ba1806abce46bf441b54a3ca4cd",8,"Research & Report","From Answer Generators to Reasoning Facilitators","Rapid integration of Large Language Models (LLMs) into educational technology risks reducing mathematics learning to answer generation. This paper reports a generative study, usability study, and 12-participant field deployment of AITutor, translating pedagogical mechanisms into UI features for junior-high students preparing for high-stakes Zhongkao exams. Mixed-methods triangulation (7,379 telemetry events, 8 observations, 10 interviews) shows students resist Socratic dialogue under time pressure, using “answer-first” shortcuts as diagnostic checkpoints. Layered worked examples, step-linked visual grounding, and metacognitive scaffolding reduce repair cost. The work proposes a Reasoning-Centered Product Loop for inspection, repair, verification, and delayed retrieval of reasoning.","From Answer Generators to Reasoning Facilitators: Designing AI Tutors for Mathematical Reasoning in High-Stakes Environments  \nHarry Feng∗ [yumingf@stanford.edu](yumingf@stanford.edu)[ ](yumingf@stanford.edu)Stanford University, EE Stanford, CA, USA  \nYuan Tian∗ [ytian24@stanford.edu](ytian24@stanford.edu)[ ](ytian24@stanford.edu)Stanford University, EALC Stanford, CA, USA  \nErica Zhao∗ [erica117@stanford.edu](erica117@stanford.edu)[ ](erica117@stanford.edu)[Stanford University](Stanford University), [MS&E](MS&E)[ ](MS&E)Stanford, CA, USA  \narXiv :2607 .0 1692v 1 [ cs .HC] 2 Jul 2026  \nAbstract  \nThe rapid integration of Large Language Models (LLMs) into educational technology threatens to reduce mathematical learning to mere answer generation. This paper presents a generative study, usability study, and 12-participant field deployment of AITutor, an interactive system that translates theoretical pedagogical mechanisms into concrete user interface features. We explore how junior-high students preparing for high-stakes exams (Zhongkao) interact with AI tutoring. Through mixed-methods triangulation (7,379 telemetry events, 8 contextual observations, 10 interviews), we reveal that students actively resist traditional Socratic dialogue under time pressure, repurposing \"answer-first\" shortcuts as vital diagnostic checkpoints. We demonstrate how features like layered worked examples, step-linked visual grounding, and metacognitive scaffolding lower the interaction cost of reasoning repair. We contribute a \"Reasoning-Centered Product Loop,\" offering actionable implications for designing AI that structurally supports the inspection, local repair, curriculum verification, and delayed retrieval of mathematical reasoning in the wild.  \nCCS Concepts  \n• Human-centered computing → HCI design and evaluation methods; • Applied computing → Interactive learning environments.  \nKeywords  \nAI Tutoring, Mathematical Reasoning, Large Language Models, Educational Technology, Cognitive Scaffolding  \nACM Reference Format:  \nHarry Feng, Yuan Tian, and Erica Zhao. 2026. From Answer Generators to Reasoning Facilitators: Designing AI Tutors for Mathematical Reasoning in High-Stakes Environments. In Proceedings of CHI Conference on Human Factors in Computing Systems (CHI ’26) . ACM, New York, NY, USA, 12 pages. [https://doi.org/10.1145/XXXXXXX.XXXXXXX](https://doi.org/10.1145/XXXXXXX.XXXXXXX)  \n∗ These authors contributed equally to this research.  \nPermission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than the author(s) must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission [and/or a fee. Request permissions from permissions@acm.org](and/or a fee. Request permissions from permissions@acm.org).  \nCHI’26, Honolulu, HI  \n© 2026 Copyright held by the owner/author(s) . Publication rights licensed to ACM. ACM ISBN 978-1-4503-XXXX-X/26/04  \n[https://doi.org/10.1145/XXXXXXX.XXXXXXX](https://doi.org/10.1145/XXXXXXX.XXXXXXX)  \n1 Introduction  \nThe fundamental goal of mathematics education is not the generation of correct answers, but the construction and refinement of robust reasoning processes. Decades of learning-science research demonstrate that durable mathematical learning requires specific pedagogical mechanisms—such as self-explanation [11], cognitive struggle [16], and metacognitive repair [27] . However, the advent of Large Language Models (LLMs) in educational technology threatens to bypass these mechanisms. By instantly delivering monolithic, highly polished solutions, modern AI tutors resolve the problem but inadvertently strip away the pedagogical scaffolding r","cbCainx5N470uC71","https://ap.wps.com/l/cbCainx5N470uC71","pdf",4198464,1,12,"English","en",105,"# Introduction\n## Research gap and motivation\n## Design and deployment of AITutor","[{\"question\":\"What problem does the paper identify with current LLM-based math tutors in high-stakes settings?\",\"answer\":\"It argues that instant, polished answer generation can bypass key learning-science mechanisms like self-explanation, cognitive struggle, and metacognitive repair, and can also create interface designs that feel like a time tax under exam pressure.\"},{\"question\":\"How does AITutor differ from traditional Socratic dialogue approaches?\",\"answer\":\"AITutor treats the interface as a balancing mechanism, supporting reasoning through UI interventions rather than forcing a binary choice between answer delivery and Socratic prompting.\"},{\"question\":\"What interface features help students repair and inspect their reasoning?\",\"answer\":\"The paper highlights layered worked examples, step-linked visual grounding to reduce multimodal coordination issues, and metacognitive scaffolding to lower the interaction cost of reasoning repair.\"}]",1784189200,30,{"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":85,"head_meta":87,"extra_data":89,"updated_unix":27},"from-answer-generators-to-reasoning-facilitators","",{"@graph":35,"@context":84},[36,53,67],{"@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/from-answer-generators-to-reasoning-facilitators/83605/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":61,"encodingFormat":60,"isAccessibleForFree":62,"interactionStatistic":63},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-16",true,{"@type":64,"interactionType":65,"userInteractionCount":4},"InteractionCounter",{"@type":66},"ViewAction",{"@type":68,"mainEntity":69},"FAQPage",[70,76,80],{"name":71,"@type":72,"acceptedAnswer":73},"What problem does the paper identify with current LLM-based math tutors in high-stakes settings?","Question",{"text":74,"@type":75},"It argues that instant, polished answer generation can bypass key learning-science mechanisms like self-explanation, cognitive struggle, and metacognitive repair, and can also create interface designs that feel like a time tax under exam pressure.","Answer",{"name":77,"@type":72,"acceptedAnswer":78},"How does AITutor differ from traditional Socratic dialogue approaches?",{"text":79,"@type":75},"AITutor treats the interface as a balancing mechanism, supporting reasoning through UI interventions rather than forcing a binary choice between answer delivery and Socratic prompting.",{"name":81,"@type":72,"acceptedAnswer":82},"What interface features help students repair and inspect their reasoning?",{"text":83,"@type":75},"The paper highlights layered worked examples, step-linked visual grounding to reduce multimodal coordination issues, and metacognitive scaffolding to lower the interaction cost of reasoning repair.","https://schema.org",{"og:url":51,"og:type":86,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":88,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":91},[92,96,100,104,109,114,119,121,126,129,133],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":93,"show_sort_weight":94,"slug":95},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":97,"show_sort_weight":98,"slug":99},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":101,"show_sort_weight":102,"slug":103},"Exam",70,"exam",{"id":105,"doc_module":4,"doc_module_name":45,"category_name":106,"show_sort_weight":107,"slug":108},5,"Comic",60,"comic",{"id":110,"doc_module":4,"doc_module_name":45,"category_name":111,"show_sort_weight":112,"slug":113},6,"Technology",50,"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":28,"slug":120},"research-report",{"id":122,"doc_module":4,"doc_module_name":45,"category_name":123,"show_sort_weight":124,"slug":125},9,"Religion & Spirituality",20,"religion-spirituality",{"id":124,"doc_module":4,"doc_module_name":45,"category_name":127,"show_sort_weight":124,"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":105,"slug":136},19,"General","general"]