[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-201206-105":59,"doc-detail-201206-en":131},{"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":124,"head_meta":126,"extra_data":128,"updated_unix":130},105,"en","supporting-calibrated-reliance-in-human-ai-collaboration-different-strategies-for-different-tasks","Supporting Calibrated Reliance in Human-AI Collaboration - Different Strategies for Different Tasks","","As AI systems increasingly support human decision making, the central problem is enabling people to judge when to trust AI predictions and when to question or override them. This work reports three controlled human-subject studies across visual reasoning and deductive logical reasoning, testing different AI support forms including predictions, probabilities, explanations, and a selective-automation policy. Results show strong task dependence: explanations help most in language-based reasoning, while predicted probabilities and benchmark policies improve visual reasoning.",{"@graph":69,"@context":123},[70,84,106],{"@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/supporting-calibrated-reliance-in-human-ai-collaboration-different-strategies-for-different-tasks/201206/",{"url":83,"name":65,"@type":85,"image":86,"author":91,"headline":65,"publisher":94,"fileFormat":97,"inLanguage":63,"description":67,"dateModified":98,"datePublished":99,"encodingFormat":97,"isAccessibleForFree":100,"interactionStatistic":101},"DigitalDocument",{"url":87,"@type":88,"width":89,"height":90},"https://docshare.wps.com/thumbnails/supporting-calibrated-reliance-in-human-ai-collaboration-different-strategies-for-different-tasks/201206.png","ImageObject",300,407,{"name":92,"@type":93},"Ophelia","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-11-01","2026-09-04",true,{"@type":102,"interactionType":103,"userInteractionCount":105},"InteractionCounter",{"@type":104},"ViewAction",17,{"@type":107,"mainEntity":108},"FAQPage",[109,115,119],{"name":110,"@type":111,"acceptedAnswer":112},"What is the main challenge in human-AI collaboration addressed in the study?","Question",{"text":113,"@type":114},"Determining when people should rely on AI predictions and when they should question or override them so that collaboration remains calibrated.","Answer",{"name":116,"@type":111,"acceptedAnswer":117},"How do different AI support strategies affect performance across tasks?",{"text":118,"@type":114},"Effectiveness is task-dependent: visual reasoning shows the strongest gains from prediction-related support, while language-based logical reasoning benefits most from LLM-generated explanations.",{"name":120,"@type":111,"acceptedAnswer":121},"Why are subjective measures like trust and confidence insufficient on their own?",{"text":122,"@type":114},"Subjective measures do not reliably predict objective decision quality, and they may not capture whether users achieve effective error recovery and properly calibrated reliance.","https://schema.org",{"og:url":83,"og:type":125,"og:title":65,"og:site_name":95,"og:description":67},"article",{"robots":127,"canonical":83},"index,follow",{"doc_id":129,"site_id":62},201206,1790381543,{"code":4,"msg":5,"data":132},{"doc_id":129,"user_id":133,"nickname":92,"user_avatar":134,"doc_module":4,"category_id":39,"category_name":40,"doc_title":65,"doc_description":67,"doc_content":135,"file_id":136,"file_url":137,"file_type":138,"file_size":139,"view_count":105,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":105,"language":140,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":141,"faqs":142,"seo_title":143,"seo_description":67,"update_tm":144,"read_time":145,"preview_page_count":4},7971461741311,"https://ap-avatar.wpscdn.com/avatar/74000253aff267980c6?x-image-process=image/resize,m_fixed,w_180,h_180&k=1779345379180704826","arXiv : 2604 .03237v2 [cs .HC] 18 Aug 2026  \nSupporting Calibrated Reliance in Human-AI Collaboration: Different Strategies for Different Tasks  \nRUTH COHEN, Bar-Ilan University, Israel LU FENG, University of Virginia, USA  \nAYALA BLOCH, Department of Psychology, Ariel University, Israel SARIT KRAUS, Bar-Ilan University, Israel  \nAs AI systems increasingly support human decision making, a central challenge is determining what information helps people recognize when to rely on AI predictions and when to question or override them. Across three controlled human-subject studies spanning abstract visual reasoning with RAVEN matrices and deductive logical reasoning with LSAT problems, we examine how different forms of AI support affect human–AI team performance. A multi-stage reveal study shows that AI predictions and explanations can affect objective accuracy and subjective confidence differently. In visual reasoning, LLM explanations do not improve accuracy beyond the predicted answer alone, and no additional support format significantly outperforms prediction-only support; predicted probabilities show the highest descriptive accuracy and error recovery, while a derived selective-automation policy provides a higher-performing reference benchmark. In language-based logical reasoning, by contrast, LLM explanations yield the highest accuracy and error recovery, outperforming expert-written explanations and probability-based support. These results show that no single support strategy is universally effective. Human–AI interfaces should instead be designed to support calibrated reliance and effective error recovery by matching the form of assistance to the task and the evidence available to users.  \nCCS Concepts: •Human-centered computing → Empirical studies inHCI; • Computing methodologies → Artificial intelligence.  \nAdditional Key Words and Phrases: Human–AI Collaboration, Explainable AI, Large Language Models  \n1 Introduction  \nAs AI systems become increasingly integrated into human decision-making workflows, a central challenge is not simply whether AI can make accurate predictions, but whether people can determine when to rely on those predictions and when to question or override them. Effective human–AI collaboration therefore requires interaction strategies that help users evaluate AI outputs and calibrate their reliance.  \nExplainable AI (XAI) has long sought to support such evaluation through methods ranging from post-hoc interpretability techniques such as LIME [16] and SHAP [9] to visual methods such as Occlusion Sensitivity (OS) [23] . More recently, Large Language Models (LLMs) have enabled fluent natural-language explanations accessible to non-experts [11, 13] . Yet explanations do not necessarily improve human–AI team performance. Human–AI teams often fail to outperform their strongest individual component [2, 18], and explanations can provide little additional benefit beyond the AI prediction or increase over-reliance and automation bias [5, 14] . Fluent but incorrect LLM rationales may further mask AI errors [17, 19] .  \nA related concern is that XAI is often evaluated using subjective measures such as trust, confidence, and perceived understanding [6], which do not always predict objective decision quality [3] . Alternative forms of support, including predicted probabilities over the answer options, may sometimes better support calibrated reliance [2, 15] . More generally, different support strategies provide  \nAuthors’ Contact Information: Ruth Cohen, Bar-Ilan University, Ramat-Gan, Israel, [ruticohen770@gmail.com](ruticohen770@gmail.com); Lu Feng, University of Virginia, Charlottesville, USA, [lu.feng@virginia.edu](lu.feng@virginia.edu); Ayala Bloch, Department of Psychology, Ariel University,  \nAriel, Israel, [ayalablo@ariel.ac.il](ayalablo@ariel.ac.il); Sarit Kraus, Bar-Ilan University, Ramat-Gan, Israel, [sarit@cs.biu.ac.il](sarit@cs.biu.ac.il).  \n2 Ruth Cohen, Lu Feng, Ayala Bloch, and Sarit Kraus  \ndifferent f","cbCaiv3hmTqMir2k","https://ap.wps.com/l/cbCaiv3hmTqMir2k","pdf",1043121,"English","# Introduction\n## Key Insights\n# The Anatomy of a Human-AI Decision","[{\"question\":\"What is the main challenge in human-AI collaboration addressed in the study?\",\"answer\":\"Determining when people should rely on AI predictions and when they should question or override them so that collaboration remains calibrated.\"},{\"question\":\"How do different AI support strategies affect performance across tasks?\",\"answer\":\"Effectiveness is task-dependent: visual reasoning shows the strongest gains from prediction-related support, while language-based logical reasoning benefits most from LLM-generated explanations.\"},{\"question\":\"Why are subjective measures like trust and confidence insufficient on their own?\",\"answer\":\"Subjective measures do not reliably predict objective decision quality, and they may not capture whether users achieve effective error recovery and properly calibrated reliance.\"}]","Supporting Calibrated Reliance in Human-AI Collaboration - Different Strategies for Different Tasks | PDF",1788527192,43]