[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-seo-515627-105":3,"detail-sidebar-cat-0-en-105":81,"doc-detail-515627-en":130},{"code":4,"msg":5,"data":6},0,"ok",{"site_id":7,"language":8,"slug":9,"title":10,"keywords":11,"description":12,"schema_data":13,"social_meta":74,"head_meta":76,"extra_data":78,"updated_unix":80},105,"en","shall-we-play-a-game-language-models-for-open-ended-wargames","Shall We Play a Game? - Language Models for Open-ended Wargames","","LLM-based social simulations can produce transcripts that appear to reflect a single behavioral signal, while the underlying system may instead perform multiple distinct functions: selecting what an actor says or does, determining what follows after an action, or covering both. This distinction is critical in open-ended wargames, where models must handle unusual actions and ambiguous consequences. The study reports a scoping review of 223 deduplicated AI-in-wargames and strategic-simulation papers, retrieved through May 1, 2026.",{"@graph":14,"@context":73},[15,34,56],{"@type":16,"itemListElement":17},"BreadcrumbList",[18,23,27,31],{"item":19,"name":20,"@type":21,"position":22},"https://docshare.wps.com","Home","ListItem",1,{"item":24,"name":25,"@type":21,"position":26},"https://docshare.wps.com/document/","Document",2,{"item":28,"name":29,"@type":21,"position":30},"https://docshare.wps.com/document/research-report/","Research & Report",3,{"item":32,"name":10,"@type":21,"position":33},"https://docshare.wps.com/document/shall-we-play-a-game-language-models-for-open-ended-wargames/515627/",4,{"url":32,"name":10,"@type":35,"image":36,"author":41,"headline":10,"publisher":44,"fileFormat":47,"inLanguage":8,"description":12,"dateModified":48,"datePublished":49,"encodingFormat":47,"isAccessibleForFree":50,"interactionStatistic":51},"DigitalDocument",{"url":37,"@type":38,"width":39,"height":40},"https://docshare.wps.com/thumbnails/shall-we-play-a-game-language-models-for-open-ended-wargames/515627.png","ImageObject",300,407,{"name":42,"@type":43},"cksankcs","Person",{"url":19,"name":45,"@type":46},"DocShare","Organization","application/pdf","2026-11-21","2026-10-06",true,{"@type":52,"interactionType":53,"userInteractionCount":55},"InteractionCounter",{"@type":54},"ViewAction",8,{"@type":57,"mainEntity":58},"FAQPage",[59,65,69],{"name":60,"@type":61,"acceptedAnswer":62},"What is the main distinction the paper emphasizes in LLM-based social simulations?","Question",{"text":63,"@type":64},"It distinguishes whether the language model is controlling player actions, adjudicating outcomes, or doing both. This separation affects how the transcript should be interpreted.","Answer",{"name":66,"@type":61,"acceptedAnswer":67},"Why is the distinction especially important in open-ended wargames?",{"text":68,"@type":64},"Open-ended wargames involve unusual actions and ambiguous consequences, so validity and safety depend on whether the model can reliably act as an adjudicator or world model, not only whether it sounds plausible.",{"name":70,"@type":61,"acceptedAnswer":71},"What does the paper contribute in its scoping review?",{"text":72,"@type":64},"It curates 223 deduplicated AI-in-wargames and strategic-simulation studies, introduces a two-axis rubric separating open-ended action from open-ended adjudication, and maps where model control sits—finding only 20 of 223 studies with both-creative roles.","https://schema.org",{"og:url":32,"og:type":75,"og:title":10,"og:site_name":45,"og:description":12},"article",{"robots":77,"canonical":32},"index,follow",{"doc_id":79,"site_id":7},515627,1791428059,{"code":4,"msg":82,"data":83},"success",[84,88,92,96,101,106,111,114,119,122,126],{"id":22,"doc_module":4,"doc_module_name":25,"category_name":85,"show_sort_weight":86,"slug":87},"Story & Novel",90,"story-novel",{"id":26,"doc_module":4,"doc_module_name":25,"category_name":89,"show_sort_weight":90,"slug":91},"Literature",80,"literature",{"id":33,"doc_module":4,"doc_module_name":25,"category_name":93,"show_sort_weight":94,"slug":95},"Exam",70,"exam",{"id":97,"doc_module":4,"doc_module_name":25,"category_name":98,"show_sort_weight":99,"slug":100},5,"Comic",60,"comic",{"id":102,"doc_module":4,"doc_module_name":25,"category_name":103,"show_sort_weight":104,"slug":105},6,"Technology",50,"technology",{"id":107,"doc_module":4,"doc_module_name":25,"category_name":108,"show_sort_weight":109,"slug":110},7,"Healthcare",40,"healthcare",{"id":55,"doc_module":4,"doc_module_name":25,"category_name":29,"show_sort_weight":112,"slug":113},30,"research-report",{"id":115,"doc_module":4,"doc_module_name":25,"category_name":116,"show_sort_weight":117,"slug":118},9,"Religion & Spirituality",20,"religion-spirituality",{"id":117,"doc_module":4,"doc_module_name":25,"category_name":120,"show_sort_weight":117,"slug":121},"World Cup","world-cup",{"id":123,"doc_module":4,"doc_module_name":25,"category_name":124,"show_sort_weight":123,"slug":125},10,"Lifestyle","lifestyle",{"id":127,"doc_module":4,"doc_module_name":25,"category_name":128,"show_sort_weight":97,"slug":129},19,"General","general",{"code":4,"msg":82,"data":131},{"doc_id":79,"user_id":132,"nickname":42,"user_avatar":133,"doc_module":4,"category_id":55,"category_name":29,"doc_title":10,"doc_description":12,"doc_content":134,"file_id":135,"file_url":136,"file_type":137,"file_size":138,"view_count":55,"is_deleted":4,"is_public":22,"is_downloadable":22,"audit_status":22,"page_count":139,"language":140,"language_code":8,"site_id":7,"html_lang":8,"table_of_contents":141,"faqs":142,"seo_title":143,"seo_description":12,"update_tm":144,"read_time":145,"preview_page_count":97,"preview_urls":146},3985747882755,"https://ap-avatar.wpscdn.com/davatar_29158cc5080c5b710cf443261637dec0","arXiv :2509 . 17 192v 3 [ cs .AI] 31 Jul 2026  \nShall We Play a Game?  \nLanguage Models for Open-ended Wargames  \nGlenn Matlin  \nIsaac Song ∗ Yixiong Hao ∗ Parv Mahajan ∗ Evan Montoya† Ryan Bard† Stuart R. Topp†  \nAnthony Wen-Ming Zang‡ Mohammed Rehan Parwani‡ Soham Shetty‡ Mark Riedl  \nCollege of Computing, Georgia Institute of Technology  \n Georgia Tech AI Safety Initiative MATS Program  \nAbstract  \nLLM-based social simulations can make a generated transcript look like a single behavioral signal, but the model behind that transcript may be doing several different jobs: choosing what an actor says or does, deciding what happens after an action, or both. The difference matters especially in open-ended wargames, where models are prized for handling unusual actions and ambiguous consequences. We report a scoping review of 223 de-duplicated AI-in-wargames and strategic-simulation papers retrieved through May 1, 2026, describing each simulation by its model-control profile: whether the language model has open-ended control over player actions, adjudication, or both. Only 20 of 223 studies (≈9%) give language models both roles. Before treating LM outputs as social simulations, researchers need to know how much creative control the model has over actions and consequences. For open-ended simulations, fidelity depends not only on whether agents behave plausibly, but also on whether language models can reliably act as adjudicators or world models.  \n1 Introduction  \nA social simulation can look plausible for the wrong reason. A transcript may show agents negotiating, escalating, compromising, or planning, but the reader still needs to know where that behavior came from. Did the model choose the participant’s action, or did a rule system constrain the result? And who decided what the action caused? Without that separation, the transcript is hard to interpret as evidence about social behavior.  \nWargames make this problem unusually clear. A serious wargame simulates conflict under uncertainty to elicit judgment about plans, assumptions, adversaries, and possible futures (Perla, 1985; Rubel, 2006; Wallman, 1995; US Army War College, 2015) . Military examples are the familiar ones, but organizations run wargaming-like simulations for diplomacy, business, public health, cybersecurity, law, and other settings where participant decisions shape later outcomes. Language models are attractive here: they can play roles, generate scenario branches, adjudicate ambiguous actions, and summarize lessons (Hogan & Brennen, 2024) . That same flexibility makes them difficult to evaluate, because once language can influence both what an actor attempts and what later becomes true in the simulated world, a fluent storyline is no longer enough.  \nThe key issue is not whether a model is “open-ended” in general but where the openness sits. A model that proposes a surprising move inside a rule-bound game creates a different kind of evidence from a model that also decides the consequences of that move. In the second setup, language holds authority over the simulated world itself, which is useful for studying improvisation, escalation, and negotiation but raises sharper validity and safety questions. Prior work already reports brittle reasoning, hallucinations, rule violations, and escalation  \n(a) Two-axis rubric with illustrative example games per quadrant.  \nPapers  \n60  \n40  \n20  \n0  \nQ-II: Analytical / Creative  \n(n=11)  \n4 in '26*  \n60  \n40  \n20  \n0  \nQ-I: Analytical / Analytical  \n(n=82)  \n13 in '26*  \n2018 2020 2022 2024 2026  \nQ-IV: Creative / Creative  \n2018 2020 2022 2024 2026  \nYear * 2026 partial (to May 1; dashed)  \n(b) Per-quadrant main-corpus paper counts over time (2018– 2026), with panels positioned as in the rubric. The dashed, lighter 2026 segment is partial (papers through May 1, 2026), not a decline. Q-III rises sharply while Q-IV (both-creative) stays rare.  \nFigure 1: Role-specific open-endedness: (a) the rubric and (b) the full 223-paper m","cbCaijlvd2Caj1Xx","https://ap.wps.com/l/cbCaijlvd2Caj1Xx","pdf",525843,49,"English","# Introduction\n## Model-control profile and role-specific openness\n# Background\n## Open-ended serious wargames","[{\"question\":\"What is the main distinction the paper emphasizes in LLM-based social simulations?\",\"answer\":\"It distinguishes whether the language model is controlling player actions, adjudicating outcomes, or doing both. This separation affects how the transcript should be interpreted.\"},{\"question\":\"Why is the distinction especially important in open-ended wargames?\",\"answer\":\"Open-ended wargames involve unusual actions and ambiguous consequences, so validity and safety depend on whether the model can reliably act as an adjudicator or world model, not only whether it sounds plausible.\"},{\"question\":\"What does the paper contribute in its scoping review?\",\"answer\":\"It curates 223 deduplicated AI-in-wargames and strategic-simulation studies, introduces a two-axis rubric separating open-ended action from open-ended adjudication, and maps where model control sits—finding only 20 of 223 studies with both-creative roles.\"}]","Shall We Play a Game? - Language Models for Open-ended Wargames | PDF",1791320281,123,[147,151,155,159,163],{"page_no":22,"url":148,"width":149,"height":150},"https://res-academy.cache.wpscdn.com/upload/kdocshare/2026-11-17/e9b96869f7907b592e363e2aeb590a2b.webp",983,1272,{"page_no":26,"url":152,"width":153,"height":154},"https://res-academy.cache.wpscdn.com/upload/kdocshare/2026-11-17/c8753f16fa0fe7c09804047318184cbf.webp",982,1270,{"page_no":30,"url":156,"width":157,"height":158},"https://res-academy.cache.wpscdn.com/upload/kdocshare/2026-11-17/b850e3c052d472f3fca6c6de0e4d45b5.webp",981,1269,{"page_no":33,"url":160,"width":161,"height":162},"https://res-academy.cache.wpscdn.com/upload/kdocshare/2026-11-17/5c528887d8c9b31fc4fa1e6f632ed736.webp",980,1268,{"page_no":97,"url":164,"width":165,"height":166},"https://res-academy.cache.wpscdn.com/upload/kdocshare/2026-11-17/80ee25c08a67ad1bfab76095627fd730.webp",979,1267]