{"id":7678,"date":"2025-10-29T05:28:00","date_gmt":"2025-10-28T20:28:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=7678"},"modified":"2025-10-25T23:10:12","modified_gmt":"2025-10-25T14:10:12","slug":"outraged-ai-large-language-models-prioritise-emotion-over-cost-in-fairness-enforcement","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=7678","title":{"rendered":"Outraged AI: Large language models prioritise emotion over cost in fairness enforcement\u00a0"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>Outraged AI: Large language models prioritise emotion over cost in fairness enforcement\u00a0<\/strong>[13.5]<br>\u6211\u3005\u306f,\u5927\u8a00\u8a9e\u30e2\u30c7\u30eb (LLM) \u304c\u611f\u60c5\u3092\u7528\u3044\u3066\u7f70\u3092\u5c0e\u3044\u3066\u3044\u308b\u3053\u3068\u3092\u793a\u3059\u3002 \u4e0d\u516c\u5e73\u306f\u3088\u308a\u5f37\u3044\u5426\u5b9a\u7684\u306a\u611f\u60c5\u3092\u3082\u305f\u3089\u3057\u3001\u3088\u308a\u591a\u304f\u306e\u7f70\u3092\u5c0e\u3044\u305f\u3002 \u5c06\u6765\u306e\u30e2\u30c7\u30eb\u3067\u306f\u3001\u4eba\u9593\u306e\u3088\u3046\u306a\u611f\u60c5\u7684\u77e5\u6027\u3092\u9054\u6210\u3059\u308b\u305f\u3081\u306b\u3001\u611f\u60c5\u3092\u6587\u8108\u306b\u654f\u611f\u306a\u63a8\u8ad6\u3068\u7d71\u5408\u3059\u3079\u304d\u3067\u3042\u308b\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2510.17880v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2510.17880v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Fri, 17 Oct 2025 08:41:36 GMT)<\/li>\n\n\n\n<li>third-party punishment (TPP) game\u3092\u7528\u3044\u305fLLM\u306e\u5206\u6790\u3002\u300cThis emotion\u2013behaviour coupling was robust and even stronger than humans across reasoning models (o3-mini, DeepSeek-R1) and an advanced foundation model (DeepSeek-V3), with the older GPT-3.5 baseline showing a weaker and less consistent effect. Analyses of the model&amp;#x27;s rationales further corroborated that elicited emotions were invoked before punitive choices (e g , references to anger in DeepSeek-R1), consistent with emotion-guided decision processes.\u300d\u3001\u300creasoning LLMs reported stronger affect to unfairness, and prioritised emotion over fairness and cost, whereas humans weighted fairness and cost more heavily75. These dissociations indicate that current LLMs have not fully internalised the human-like cost\u2013benefit calculus that tempers norm enforcement.\u300d\u306a\u3069\u9762\u767d\u3044\u7d50\u679c\u304c\u591a\u3044\u3002LLM\/LRM\u3092\u9ad8\u5ea6\u306a\u5206\u91ce\u306b\u4f7f\u3063\u3066\u3044\u304f\u306b\u3042\u305f\u3063\u3066\u306f\u4eba\u9593\u3068\u306e\u5dee\u7570\u304c\u3042\u308b\u3053\u3068\u3092\u5927\u524d\u63d0\u3068\u3057\u3066\u6ce8\u610f\u6df1\u3044\u8a55\u4fa1\u304c\u5fc5\u8981\u306a\u3093\u3060\u308d\u3046\u3068\u601d\u3046\u3002<\/li>\n\n\n\n<li>\u300c<a href=\"https:\/\/devneko.jp\/wordpress\/?p=7045\">Corrupted by Reasoning: Reasoning Language Models Become Free-Riders in Public Goods Games, How large language models judge and influence human cooperation \u2013 arXiv\u6700\u65b0\u8ad6\u6587\u306e\u7d39\u4ecb<\/a>\u300d\u3067\u3082\u601d\u3063\u305f\u304c\u3053\u306e\u624b\u306e\u7814\u7a76\u306f\u3068\u3066\u3082\u8208\u5473\u6df1\u3044\u3002<\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[2],"tags":[223],"class_list":["post-7678","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-llm"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/7678","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=7678"}],"version-history":[{"count":1,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/7678\/revisions"}],"predecessor-version":[{"id":7679,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/7678\/revisions\/7679"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=7678"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=7678"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=7678"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}