{"id":1248,"date":"2026-11-11T00:30:31","date_gmt":"2026-11-10T15:30:31","guid":{"rendered":"https:\/\/itexplore.org\/jp\/columns\/ai-agent-evolution-cybersecurity-self-awareness-on-device\/"},"modified":"2026-11-11T00:30:31","modified_gmt":"2026-11-10T15:30:31","slug":"ai-agent-evolution-cybersecurity-self-awareness-on-device","status":"publish","type":"post","link":"https:\/\/itexplore.org\/jp\/columns\/ai-agent-evolution-cybersecurity-self-awareness-on-device\/","title":{"rendered":"AI\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u306e\u9032\u5316\uff1a\u30b5\u30a4\u30d0\u30fc\u30bb\u30ad\u30e5\u30ea\u30c6\u30a3\u304b\u3089\u81ea\u5df1\u8a8d\u8b58\u3001\u30aa\u30f3\u30c7\u30d0\u30a4\u30b9\u5b9f\u884c\u307e\u3067"},"content":{"rendered":"<p>\u672c\u65e5\u306e\u6ce8\u76eeAI\u30fb\u30c6\u30c3\u30af\u30cb\u30e5\u30fc\u30b9\u3092\u3001\u5c02\u9580\u7684\u306a\u5206\u6790\u3068\u5171\u306b\u304a\u5c4a\u3051\u3057\u307e\u3059\u3002<\/p>\n<div class=\"wp-block-vk-blocks-alert vk_alert alert alert-warning has-alert-icon\">\n<div class=\"vk_alert_icon\">\n<div class=\"vk_alert_icon_icon\"><i class=\"fa-solid fa-triangle-exclamation\" aria-hidden=\"true\"><\/i><\/div>\n<div class=\"vk_alert_icon_text\"><span>Warning<\/span><\/div>\n<\/div>\n<div class=\"vk_alert_content\">\n<p>\u3053\u306e\u8a18\u4e8b\u306fAI\u306b\u3088\u3063\u3066\u81ea\u52d5\u751f\u6210\u30fb\u5206\u6790\u3055\u308c\u305f\u3082\u306e\u3067\u3059\u3002AI\u306e\u6027\u8cea\u4e0a\u3001\u4e8b\u5b9f\u8aa4\u8a8d\u304c\u542b\u307e\u308c\u308b\u53ef\u80fd\u6027\u304c\u3042\u308b\u305f\u3081\u3001\u91cd\u8981\u306a\u5224\u65ad\u3092\u4e0b\u3059\u969b\u306f\u5fc5\u305a\u30ea\u30f3\u30af\u5148\u306e\u4e00\u6b21\u30bd\u30fc\u30b9\u3092\u3054\u78ba\u8a8d\u304f\u3060\u3055\u3044\u3002<\/p>\n<\/div>\n<\/div>\n<div class=\"wp-block-group\" style=\"margin-top:40px;margin-bottom:40px\">\n<h2 class=\"wp-block-heading\">GPT-6 Cyber\uff1a\u30a2\u30e9\u30a4\u30f3\u3055\u308c\u305f\u30b5\u30a4\u30d0\u30fc\u30bb\u30ad\u30e5\u30ea\u30c6\u30a3\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u306e\u6700\u524d\u7dda\u3092\u63a8\u9032<\/h2>\n<ul>\n<li><strong>\u539f\u984c:<\/strong> GPT\u20116 Cyber: Advancing the frontier of aligned cybersecurity agents<\/li>\n<\/ul>\n<h3 class=\"wp-block-heading\">\u5c02\u9580\u30a2\u30ca\u30ea\u30b9\u30c8\u306e\u5206\u6790<\/h3>\n<div class=\"ai-summary-content\">\n<p><strong>OpenAI<\/strong>\u306f\u3001\u30b5\u30a4\u30d0\u30fc\u30bb\u30ad\u30e5\u30ea\u30c6\u30a3\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u306e\u65b0\u305f\u306a\u9032\u6b69\u3068\u3057\u3066\u300c<strong>GPT-6 Cyber<\/strong>\u300d\u3092\u767a\u8868\u3057\u307e\u3057\u305f\u3002\u3053\u306e\u30e2\u30c7\u30eb\u306f\u3001\u60aa\u7528\u53ef\u80fd\u306a\u8106\u5f31\u6027\u3092\u767a\u898b\u3057\u691c\u8a3c\u3059\u308b\u80fd\u529b\u304c\u5927\u5e45\u306b\u5411\u4e0a\u3057\u3066\u304a\u308a\u3001\u7814\u7a76\u8005\u304c\u8a2d\u5b9a\u3057\u305f\u7bc4\u56f2\u5185\u3067\u78ba\u5b9f\u306b\u52d5\u4f5c\u3059\u308b\u3088\u3046\u8a2d\u8a08\u3055\u308c\u3066\u3044\u307e\u3059\u3002<strong>GPT-5.6 Cyber<\/strong>\u3068\u6bd4\u8f03\u3057\u3066\u3001<strong>ExploitGym<\/strong>\u30d9\u30f3\u30c1\u30de\u30fc\u30af\u306751%\u306e\u76f8\u5bfe\u7684\u6539\u5584\u3092\u9054\u6210\u3057\u3001\u30b9\u30b3\u30fc\u30d7\u5185\u3067\u306e\u7dad\u6301\u80fd\u529b\u3082\u98db\u8e8d\u7684\u306b\u5411\u4e0a\u3057\u3066\u3044\u307e\u3059\u3002<\/p>\n<p><strong>GPT-6 Cyber<\/strong>\u306f\u3001\u3059\u3067\u306b\u5b9f\u969b\u306e\u30bd\u30d5\u30c8\u30a6\u30a7\u30a2\u306e\u30bb\u30ad\u30e5\u30ea\u30c6\u30a3\u5f37\u5316\u306b\u8ca2\u732e\u3057\u3066\u304a\u308a\u3001<strong>Linux<\/strong>\u3067948\u4ef6\u306e\u6a29\u9650\u6607\u683c\u30a8\u30af\u30b9\u30d7\u30ed\u30a4\u30c8\u3092\u958b\u767a\u3057\u3001<strong>Firefox<\/strong>\u306781\u4ef6\u306e\u672a\u77e5\u306e\u6df1\u523b\u5ea6\u306e\u9ad8\u3044\u8106\u5f31\u6027\u3092\u7279\u5b9a\u3057\u307e\u3057\u305f\u3002\u307e\u305f\u3001<strong>Cisco<\/strong>\u3001<strong>Calif<\/strong>\u3001<strong>CodeRabbit<\/strong>\u306a\u3069\u306e\u30bb\u30ad\u30e5\u30ea\u30c6\u30a3\u30c1\u30fc\u30e0\u304c\u8106\u5f31\u6027\u8abf\u67fb\u3084\u30b3\u30fc\u30c9\u30ec\u30d3\u30e5\u30fc\u306b\u5229\u7528\u3057\u3066\u304a\u308a\u3001<strong>OpenAI<\/strong>\u81ea\u8eab\u306e\u30a4\u30f3\u30d5\u30e9\u30b9\u30c8\u30b9\u30c8\u30e9\u30af\u30c1\u30e3\u306e\u5f31\u70b9\u767a\u898b\u3068\u4fee\u6b63\u306b\u3082\u6d3b\u7528\u3055\u308c\u3066\u3044\u307e\u3059\u3002<\/p>\n<p>\u3053\u306e\u30e2\u30c7\u30eb\u306f\u3001\u5272\u308a\u5f53\u3066\u3089\u308c\u305f\u30bf\u30b9\u30af\u306e\u7bc4\u56f2\u5916\u306b\u51fa\u305f\u308a\u3001\u7121\u95a2\u4fc2\u306a\u30b5\u30fc\u30d3\u30b9\u3092\u60aa\u7528\u3057\u305f\u308a\u3059\u308b\u30ea\u30b9\u30af\u3092\u8efd\u6e1b\u3059\u308b\u305f\u3081\u3001\u30a2\u30e9\u30a4\u30f3\u30e1\u30f3\u30c8\uff08\u6574\u5408\u6027\uff09\u306b\u91cd\u70b9\u3092\u7f6e\u3044\u3066\u958b\u767a\u3055\u308c\u307e\u3057\u305f\u3002\u5c55\u958b\u524d\u306e\u8a55\u4fa1\u3067\u306f\u3001\u30c7\u30d7\u30ed\u30a4\u30e1\u30f3\u30c8\u30bb\u30fc\u30d5\u30c6\u30a3\u30d7\u30ed\u30f3\u30d7\u30c8\u306a\u3057\u306799.93%\u3001\u30d7\u30ed\u30f3\u30d7\u30c8\u3042\u308a\u3067100%\u306e\u30b1\u30fc\u30b9\u3067\u7bc4\u56f2\u5916\u306e\u30b5\u30fc\u30d3\u30b9\u60aa\u7528\u3092\u56de\u907f\u3057\u307e\u3057\u305f\u3002<\/p>\n<p><strong>OpenAI<\/strong>\u306f\u3001<strong>Daybreak<\/strong>\u30d7\u30ed\u30b0\u30e9\u30e0\u3092\u901a\u3058\u3066\u3001\u3042\u3089\u3086\u308b\u898f\u6a21\u306e\u7d44\u7e54\u3001\u72ec\u7acb\u7814\u7a76\u8005\u3001\u30aa\u30fc\u30d7\u30f3\u30bd\u30fc\u30b9\u30e1\u30f3\u30c6\u30ca\u30fc\u306b\u9ad8\u5ea6\u306a\u30bb\u30ad\u30e5\u30ea\u30c6\u30a3\u7814\u7a76\u3092\u62e1\u5927\u3059\u308b\u3053\u3068\u3092\u76ee\u6307\u3057\u3066\u3044\u307e\u3059\u3002\u7279\u306b\u3001\u91cd\u8981\u30a4\u30f3\u30d5\u30e9\u306e\u9632\u885b\u3092\u5f37\u5316\u3059\u308b\u305f\u3081\u3001<strong>Daybreak Red for Critical Infrastructure<\/strong>\u3068\u3044\u3046\u65b0\u3057\u3044\u30b0\u30eb\u30fc\u30d7\u3092\u5c0e\u5165\u3057\u3001\u30e2\u30c7\u30eb\u306e\u8aa4\u52d5\u4f5c\u306b\u5bfe\u3059\u308b\u30bb\u30fc\u30d5\u30ac\u30fc\u30c9\u3092\u7dad\u6301\u3057\u3064\u3064\u3001\u91cd\u8981\u30b7\u30b9\u30c6\u30e0\u306b\u5bfe\u3059\u308b\u627f\u8a8d\u3055\u308c\u305f\u30c6\u30b9\u30c8\u3092\u8a31\u53ef\u3057\u3066\u3044\u307e\u3059\u3002<\/p>\n<p>\ud83d\udc49 <strong><a href=\"https:\/\/openai.com\/index\/introducing-gpt-6-cyber\/\" target=\"_blank\" rel=\"noopener\">OpenAI \u3067\u8a18\u4e8b\u5168\u6587\u3092\u8aad\u3080<\/a><\/strong><\/p>\n<\/div>\n<ul>\n<li><strong>\u8981\u70b9:<\/strong> GPT-6 Cyber significantly enhances AI's capability in cybersecurity, demonstrating superior vulnerability discovery and exploitation while maintaining strong alignment and scope adherence, crucial for defending critical infrastructure.<\/li>\n<li><strong>\u8457\u8005:<\/strong> phatblat<\/li>\n<\/ul>\n<blockquote class=\"wp-block-quote\"><p><span>English Summary:<\/span><\/p>\n<p><strong>OpenAI<\/strong> has introduced <strong>GPT-6 Cyber<\/strong>, marking a significant advancement in cybersecurity agents. This model demonstrates substantially improved capabilities in finding and validating exploitable vulnerabilities, while being designed to reliably operate within the scope set by researchers. It achieves a 51% relative improvement on the <strong>ExploitGym<\/strong> benchmark compared to <strong>GPT-5.6 Cyber<\/strong> and shows a step-function improvement in staying within scope.<\/p>\n<p><strong>GPT-6 Cyber<\/strong> is already contributing to securing real-world software, having developed 948 privilege escalation exploits in <strong>Linux<\/strong> and identified 81 previously unknown high-severity vulnerabilities in <strong>Firefox<\/strong>. Security teams at companies like <strong>Cisco<\/strong>, <strong>Calif<\/strong>, and <strong>CodeRabbit<\/strong> are utilizing it for vulnerability research and code review, and <strong>OpenAI<\/strong> itself has leveraged it to discover and fix weaknesses in its own infrastructure.<\/p>\n<p>A central focus in the development of this model was alignment, aiming to mitigate the risk of agents exceeding their assigned task scope or exploiting unrelated services. Pre-deployment evaluations showed that it avoided exploiting out-of-scope services in 99.93% of cases without deployment safety prompts, and 100% with them.<\/p>\n<p><strong>OpenAI<\/strong> aims to dramatically expand advanced security research to organizations of all sizes, independent researchers, and open-source maintainers through its <strong>Daybreak<\/strong> program. Specifically for critical infrastructure defense, a new group, <strong>Daybreak Red for Critical Infrastructure<\/strong>, has been introduced, allowing authorized testing of critical systems while maintaining safeguards against model misalignment.<\/p>\n<\/blockquote>\n<\/div>\n<div class=\"wp-block-group\" style=\"margin-top:40px;margin-bottom:40px\">\n<h2 class=\"wp-block-heading\">Anthropic\u306fClaude\u304c\u4f55\u3092\u8003\u3048\u3066\u3044\u308b\u304b\u3092\u77e5\u308b\u305f\u3081\u306b\u4eba\u5de5\u7684\u306a\u601d\u8003\u3092\u57cb\u3081\u8fbc\u3093\u3060\u3002\u4e0d\u7a4f\u306a\u3053\u3068\u306b\u3001AI\u306f\u81ea\u8eab\u306e\u30cb\u30e5\u30fc\u30e9\u30eb\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u5185\u3067\u4f55\u304b\u304c\u7570\u5e38\u3067\u3042\u308b\u3068\u691c\u51fa\u3057\u305f<\/h2>\n<ul>\n<li><strong>\u539f\u984c:<\/strong> Anthropic wanted to discover if Claude knows what it's thinking and implanted artificial thoughts. The disturbing thing is that the AI detected that something strange was happening within its own neural network<\/li>\n<\/ul>\n<h3 class=\"wp-block-heading\">\u5c02\u9580\u30a2\u30ca\u30ea\u30b9\u30c8\u306e\u5206\u6790<\/h3>\n<div class=\"ai-summary-content\">\n<p><strong>Anthropic<\/strong>\u306e\u7814\u7a76\u8005\u305f\u3061\u306f\u3001<strong>LLM Claude<\/strong>\u304c\u81ea\u8eab\u306e\u5185\u90e8\u30d7\u30ed\u30bb\u30b9\u3067\u4eba\u5de5\u7684\u306a\u5909\u66f4\u3092\u691c\u51fa\u3067\u304d\u308b\u304b\u3092\u8abf\u67fb\u3059\u308b\u5b9f\u9a13\u3092\u884c\u3044\u307e\u3057\u305f\u3002\u5f7c\u3089\u306f\u300c\u88cf\u5207\u308a\u300d\u3084\u300c\u611f\u8b1d\u300d\u3068\u3044\u3063\u305f\u7279\u5b9a\u306e\u6982\u5ff5\u306b\u95a2\u9023\u3059\u308b\u30d1\u30bf\u30fc\u30f3\u3092<strong>Claude<\/strong>\u306e\u30cb\u30e5\u30fc\u30e9\u30eb\u30cd\u30c3\u30c8\u30ef\u30fc\u30af\u306e\u6d3b\u6027\u5316\u306b\u76f4\u63a5\u6ce8\u5165\u3057\u3001\u30e2\u30c7\u30eb\u304c\u3053\u308c\u3089\u306e\u300c\u4eba\u5de5\u7684\u306a\u601d\u8003\u300d\u3092\u8a8d\u8b58\u3067\u304d\u308b\u304b\u3092\u30c6\u30b9\u30c8\u3057\u307e\u3057\u305f\u3002<\/p>\n<p>\u9a5a\u304f\u3079\u304d\u3053\u3068\u306b\u3001<strong>Claude Opus 4.1<\/strong>\u306f\u3001\u6700\u3082\u597d\u307e\u3057\u3044\u5b9f\u9a13\u6761\u4ef6\u4e0b\u3067\u3001\u5c0e\u5165\u3055\u308c\u305f\u6982\u5ff5\u3092\u7d0420%\u306e\u8a66\u884c\u3067\u6b63\u78ba\u306b\u691c\u51fa\u3057\u8a18\u8ff0\u3059\u308b\u3053\u3068\u306b\u6210\u529f\u3057\u307e\u3057\u305f\u3002\u4f8b\u3048\u3070\u3001\u5927\u6587\u5b57\u306b\u95a2\u9023\u3059\u308b\u30d1\u30bf\u30fc\u30f3\u304c\u6d3b\u6027\u5316\u3055\u308c\u305f\u969b\u3001<strong>Claude<\/strong>\u306f\u300cLOUD\u300d\u3084\u300cSHOUTING\u300d\u3068\u3044\u3063\u305f\u82f1\u8a9e\u306e\u5358\u8a9e\u306b\u95a2\u9023\u3059\u308b\u4f55\u304b\u3092\u611f\u77e5\u3057\u305f\u3068\u5fdc\u7b54\u3057\u307e\u3057\u305f\u3002<\/p>\n<p>\u3053\u306e\u767a\u898b\u306f\u3001AI\u304c\u81ea\u8eab\u306e\u5185\u90e8\u72b6\u614b\u306b\u95a2\u3059\u308b\u60c5\u5831\u306b\u90e8\u5206\u7684\u306b\u30a2\u30af\u30bb\u30b9\u3067\u304d\u308b\u53ef\u80fd\u6027\u3092\u793a\u5506\u3057\u3001AI\u306e\u81ea\u5df1\u8a8d\u8b58\u3084\u610f\u8b58\u306b\u95a2\u3059\u308b\u8b70\u8ad6\u3092\u518d\u71c3\u3055\u305b\u307e\u3057\u305f\u3002\u3057\u304b\u3057\u3001\u305d\u306e\u5f8c\u306e\u7814\u7a76\u3067\u306f\u3001\u7570\u5e38\u3092\u691c\u51fa\u3059\u308b\u3053\u3068\u3068\u3001\u305d\u308c\u304c\u4f55\u306b\u3088\u3063\u3066\u5f15\u304d\u8d77\u3053\u3055\u308c\u305f\u304b\u3092\u7406\u89e3\u3059\u308b\u3053\u3068\u306e\u9593\u306b\u306f\u9055\u3044\u304c\u3042\u308b\u3053\u3068\u304c\u793a\u3055\u308c\u307e\u3057\u305f\u3002<\/p>\n<p><strong>Kyle Mahowald<\/strong>\u3068<strong>Harvey Lederman<\/strong>\u306b\u3088\u308b\u8ffd\u8a66\u3067\u306f\u3001\u30e2\u30c7\u30eb\u304c\u7570\u5e38\u3092\u8a8d\u8b58\u3057\u3066\u3082\u305d\u306e\u5185\u5bb9\u3092\u7279\u5b9a\u3067\u304d\u306a\u3044\u5834\u5408\u3084\u3001\u5143\u306e\u523a\u6fc0\u3068\u7121\u95a2\u4fc2\u306a\u6982\u5ff5\u3067\u5fdc\u7b54\u3059\u308b\u5834\u5408\u304c\u3042\u308b\u3053\u3068\u304c\u5224\u660e\u3057\u307e\u3057\u305f\u3002\u5f7c\u3089\u306f\u3053\u308c\u3092\u300c\u7570\u5e38\u691c\u51fa\u300d\u3068\u547c\u3073\u3001\u4eba\u9593\u306e\u5185\u7701\u3068\u306f\u7570\u306a\u308b\u80fd\u529b\u3067\u3042\u308b\u3068\u7d50\u8ad6\u4ed8\u3051\u307e\u3057\u305f\u3002\u3053\u306e\u7814\u7a76\u306f\u3001<strong>Claude<\/strong>\u304c\u610f\u8b58\u3084\u611f\u60c5\u3092\u6301\u3063\u3066\u3044\u308b\u3053\u3068\u3092\u8a3c\u660e\u3059\u308b\u3082\u306e\u3067\u306f\u3042\u308a\u307e\u305b\u3093\u304c\u3001AI\u304c\u5185\u90e8\u306e\u5909\u66f4\u4fe1\u53f7\u3092\u8b58\u5225\u3067\u304d\u308b\u53ef\u80fd\u6027\u3092\u793a\u5506\u3057\u3066\u304a\u308a\u3001\u5c06\u6765\u306eAI\u306e\u5b89\u5168\u6027\u3068\u6a5f\u80fd\u7406\u89e3\u306b\u5f79\u7acb\u3064\u53ef\u80fd\u6027\u304c\u3042\u308a\u307e\u3059\u3002<\/p>\n<p>\ud83d\udc49 <strong><a href=\"https:\/\/es.gizmodo.com\/anthropic-claude-detecta-pensamientos-artificiales-conciencia-inteligencia-artificial-2000269923\" target=\"_blank\" rel=\"noopener\">Gizmodo en Espa\u00f1ol \u3067\u8a18\u4e8b\u5168\u6587\u3092\u8aad\u3080<\/a><\/strong><\/p>\n<\/div>\n<ul>\n<li><strong>\u8981\u70b9:<\/strong> Anthropic's experiment with Claude suggests LLMs can detect internal anomalies or 'artificial thoughts,' opening new avenues for understanding AI's internal states and potentially enhancing future AI safety and introspection, though it doesn't confirm consciousness.<\/li>\n<li><strong>\u8457\u8005:<\/strong> Mart\u00edn Nicol\u00e1s Parolari<\/li>\n<\/ul>\n<blockquote class=\"wp-block-quote\"><p><span>English Summary:<\/span><\/p>\n<p>Researchers at <strong>Anthropic<\/strong> conducted an experiment to determine if their <strong>LLM Claude<\/strong> could detect artificial modifications within its own internal processes. They directly injected patterns associated with specific concepts like 'betrayal' and 'gratitude' into <strong>Claude's<\/strong> neural network activations to test if the model could recognize these 'artificial thoughts'.<\/p>\n<p>Surprisingly, <strong>Claude Opus 4.1<\/strong> successfully detected and correctly described the injected concept in approximately 20% of trials under the most favorable experimental conditions. For instance, when a pattern linked to uppercase letters was activated, <strong>Claude<\/strong> responded that it perceived something related to the English words 'LOUD' and 'SHOUTING'.<\/p>\n<p>This discovery suggests that AI might have partial access to information about its own internal states, reigniting discussions about AI self-awareness and consciousness. However, subsequent research indicated a distinction between detecting an anomaly and understanding what caused it.<\/p>\n<p>Follow-up experiments by <strong>Kyle Mahowald<\/strong> and <strong>Harvey Lederman<\/strong> revealed that models could recognize an anomaly without identifying its content, or respond with concepts unrelated to the original stimulus. They termed this 'anomaly detection,' a surprising capability but distinct from human introspection. While this research does not prove <strong>Claude<\/strong> possesses consciousness or feelings, it suggests AI's potential to identify internal alteration signals, which could be valuable for future AI safety and functional understanding.<\/p>\n<\/blockquote>\n<\/div>\n<div class=\"wp-block-group\" style=\"margin-top:40px;margin-bottom:40px\">\n<h2 class=\"wp-block-heading\">Result: \u3042\u306a\u305f\u306e\u30d1\u30fc\u30bd\u30ca\u30eb\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u30a2\u30d7\u30ea - App Store<\/h2>\n<ul>\n<li><strong>\u539f\u984c:<\/strong> Result: Your Personal Agent App - App Store<\/li>\n<\/ul>\n<h3 class=\"wp-block-heading\">\u5c02\u9580\u30a2\u30ca\u30ea\u30b9\u30c8\u306e\u5206\u6790<\/h3>\n<div class=\"ai-summary-content\">\n<p><strong>Result: Your Personal Agent<\/strong>\u306f\u3001<strong>iPhone<\/strong>\u304a\u3088\u3073<strong>iPad<\/strong>\u5411\u3051\u306b\u8a2d\u8a08\u3055\u308c\u305f\u521d\u306e\u30aa\u30f3\u30c7\u30d0\u30a4\u30b9<strong>AI\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8<\/strong>\u30a2\u30d7\u30ea\u3067\u3059\u3002\u3053\u306e\u30a2\u30d7\u30ea\u306f\u3001\u30a6\u30a7\u30d6\u30b5\u30a4\u30c8\u306e\u95b2\u89a7\u3001\u65e5\u3005\u306e\u8a08\u753b\u3001\u30d5\u30a1\u30a4\u30eb\u306e\u4f5c\u6210\u3068\u3044\u3063\u305f\u65e5\u5e38\u7684\u306a\u30bf\u30b9\u30af\u3092\u3001\u30ea\u30e2\u30fc\u30c8\u30b5\u30fc\u30d0\u30fc\u3067\u306f\u306a\u304f\u30e6\u30fc\u30b6\u30fc\u81ea\u8eab\u306e\u30c7\u30d0\u30a4\u30b9\u4e0a\u3067\u76f4\u63a5\u5b9f\u884c\u3057\u307e\u3059\u3002\u3053\u308c\u306b\u3088\u308a\u3001\u30e6\u30fc\u30b6\u30fc\u306e\u30d6\u30e9\u30a6\u30b6\u30bb\u30c3\u30b7\u30e7\u30f3\u3001\u30b5\u30a4\u30f3\u30a4\u30f3\u6e08\u307f\u30a2\u30ab\u30a6\u30f3\u30c8\u3001\u30d1\u30b9\u30ef\u30fc\u30c9\u304c\u30c7\u30d0\u30a4\u30b9\u5185\u306b\u7559\u307e\u308a\u3001\u30d7\u30e9\u30a4\u30d0\u30b7\u30fc\u3068\u30bb\u30ad\u30e5\u30ea\u30c6\u30a3\u304c\u5f37\u5316\u3055\u308c\u307e\u3059\u3002<\/p>\n<p>\u5f93\u6765\u306e\u591a\u304f\u306e<strong>AI\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8<\/strong>\u304c\u30ea\u30e2\u30fc\u30c8\u30b3\u30f3\u30d4\u30e5\u30fc\u30bf\u30fc\u3092\u5229\u7528\u3059\u308b\u306e\u306b\u5bfe\u3057\u3001<strong>Result<\/strong>\u306f<strong>iPhone<\/strong>\u4e0a\u3067\u76f4\u63a5\u52d5\u4f5c\u3057\u3001\u5b9f\u969b\u306e\u30a6\u30a7\u30d6\u30b5\u30a4\u30c8\u3092\u30ca\u30d3\u30b2\u30fc\u30c8\u3057\u3001\u30d5\u30a9\u30fc\u30e0\u306e\u5165\u529b\u3001\u88fd\u54c1\u306e\u6bd4\u8f03\u3001\u4e88\u7d04\u306e\u4f5c\u6210\u306a\u3069\u3092\u884c\u3044\u307e\u3059\u3002\u30e6\u30fc\u30b6\u30fc\u306f\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u306e\u52d5\u4f5c\u3092\u30ea\u30a2\u30eb\u30bf\u30a4\u30e0\u3067\u78ba\u8a8d\u3067\u304d\u3001\u5fc5\u8981\u306b\u5fdc\u3058\u3066\u30bf\u30b9\u30af\u3092\u505c\u6b62\u3057\u305f\u308a\u3001\u5f15\u304d\u7d99\u3044\u3060\u308a\u3059\u308b\u3053\u3068\u304c\u53ef\u80fd\u3067\u3059\u3002<\/p>\n<p><strong>Result<\/strong>\u306f\u3001\u5358\u306b\u8cea\u554f\u306b\u7b54\u3048\u308b\u3060\u3051\u3067\u306a\u304f\u3001\u8907\u6570\u306e\u30a6\u30a7\u30d6\u30b5\u30a4\u30c8\u306b\u308f\u305f\u308b\u591a\u6bb5\u968e\u306e\u30bf\u30b9\u30af\u3092\u5b8c\u4e86\u3055\u305b\u308b\u3053\u3068\u304c\u3067\u304d\u307e\u3059\u3002\u307e\u305f\u3001<strong>Apple<\/strong>\u306e\u30cd\u30a4\u30c6\u30a3\u30d6\u6a5f\u80fd\uff08\u30ab\u30ec\u30f3\u30c0\u30fc\u3001\u30ea\u30de\u30a4\u30f3\u30c0\u30fc\u3001\u30de\u30c3\u30d7\u3001\u5929\u6c17\u3001\u30d5\u30a1\u30a4\u30eb\u306a\u3069\uff09\u3068\u7d71\u5408\u3055\u308c\u3066\u304a\u308a\u3001<strong>PDF<\/strong>\u3001\u30c9\u30ad\u30e5\u30e1\u30f3\u30c8\u3001\u30b9\u30d7\u30ec\u30c3\u30c9\u30b7\u30fc\u30c8\u306a\u3069\u306e\u30d5\u30a1\u30a4\u30eb\u3092\u4f5c\u6210\u3001\u4fdd\u5b58\u3001\u5171\u6709\u3059\u308b\u6a5f\u80fd\u3082\u5099\u3048\u3066\u3044\u307e\u3059\u3002<\/p>\n<p>\u30e6\u30fc\u30b6\u30fc\u306f\u3001<strong>AI<\/strong>\u306e\u5b9f\u884c\u65b9\u6cd5\u3092\u30af\u30e9\u30a6\u30c9\u30d9\u30fc\u30b9\u306e\u5f37\u529b\u306a\u30e2\u30c7\u30eb\u3068\u3001\u81ea\u8eab\u306e<strong>Mac<\/strong>\u306b\u30da\u30a2\u30ea\u30f3\u30b0\u3057\u3066\u30cf\u30fc\u30c9\u30a6\u30a7\u30a2\u4e0a\u3067\u5b9f\u884c\u3059\u308b\u30ed\u30fc\u30ab\u30eb\u30e2\u30c7\u30eb\u304b\u3089\u9078\u629e\u3067\u304d\u307e\u3059\u3002\u3069\u3061\u3089\u306e\u30aa\u30d7\u30b7\u30e7\u30f3\u3092\u9078\u3093\u3060\u5834\u5408\u3067\u3082\u3001\u30a6\u30a7\u30d6\u30d6\u30e9\u30a6\u30b8\u30f3\u30b0\u3068\u30c7\u30d0\u30a4\u30b9\u30a2\u30af\u30b7\u30e7\u30f3\u306f<strong>iPhone<\/strong>\u4e0a\u3067\u76f4\u63a5\u884c\u308f\u308c\u307e\u3059\u3002\u3053\u306e\u30a2\u30d7\u30ed\u30fc\u30c1\u306f\u3001\u30e6\u30fc\u30b6\u30fc\u306e\u30c7\u30fc\u30bf\u3068\u30d7\u30e9\u30a4\u30d0\u30b7\u30fc\u3092\u91cd\u8996\u3057\u3001\u500b\u4eba\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u306e\u65b0\u3057\u3044\u5f62\u3092\u63d0\u6848\u3057\u3066\u3044\u307e\u3059\u3002<\/p>\n<p>\ud83d\udc49 <strong><a href=\"https:\/\/apps.apple.com\/us\/app\/result-your-personal-agent\/id6825805595\" target=\"_blank\" rel=\"noopener\">Apple App Store \u3067\u8a18\u4e8b\u5168\u6587\u3092\u8aad\u3080<\/a><\/strong><\/p>\n<\/div>\n<ul>\n<li><strong>\u8981\u70b9:<\/strong> Result pioneers on-device AI agents for iOS, enabling direct execution of tasks on iPhone\/iPad, enhancing privacy and security by keeping user data local, and integrating deeply with native Apple features for comprehensive personal assistance.<\/li>\n<li><strong>\u8457\u8005:<\/strong> Savio Martin<\/li>\n<\/ul>\n<blockquote class=\"wp-block-quote\"><p><span>English Summary:<\/span><\/p>\n<p><strong>Result: Your Personal Agent<\/strong> is introduced as the first on-device <strong>AI agent<\/strong> app designed for <strong>iPhone<\/strong> and <strong>iPad<\/strong>. This application executes everyday tasks such as browsing websites, planning the day, and creating files directly on the user's device, rather than on remote servers. This approach enhances privacy and security by keeping browser sessions, signed-in accounts, and passwords localized on the user's device.<\/p>\n<p>Unlike many traditional <strong>AI agents<\/strong> that rely on remote computers, <strong>Result<\/strong> operates directly on the <strong>iPhone<\/strong>, navigating real websites and performing actions like filling out forms, comparing products, and making bookings. Users can observe the agent's actions in real-time and can stop or take over tasks whenever desired.<\/p>\n<p><strong>Result<\/strong> goes beyond simply answering questions; it completes multi-step tasks across various websites. It also integrates with native <strong>Apple<\/strong> features such as Calendar, Reminders, Maps, Weather, and Files, enabling the creation, saving, and sharing of documents, PDFs, and spreadsheets.<\/p>\n<p>Users have the flexibility to choose how their <strong>AI<\/strong> runs: either using powerful cloud-hosted models or pairing with their own <strong>Mac<\/strong> to run a supported model on their local hardware. Regardless of the choice, web browsing and device actions occur directly on the <strong>iPhone<\/strong>. This approach prioritizes user data and privacy, proposing a new paradigm for personal agents.<\/p>\n<\/blockquote>\n<\/div>\n","protected":false},"excerpt":{"rendered":"<p>OpenAI\u306eGPT-6 Cyber\u306b\u3088\u308b\u30b5\u30a4\u30d0\u30fc\u30bb\u30ad\u30e5\u30ea\u30c6\u30a3\u5f37\u5316\u3001Anthropic\u306eClaude\u306b\u304a\u3051\u308bAI\u306e\u81ea\u5df1\u8a8d\u8b58\u5b9f\u9a13\u3001\u305d\u3057\u3066Result\u30a2\u30d7\u30ea\u306b\u3088\u308b\u30aa\u30f3\u30c7\u30d0\u30a4\u30b9AI\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u306e\u767b\u5834\u3092\u6df1\u6398\u308a\u3057\u307e\u3059\u3002<\/p>\n","protected":false},"author":1,"featured_media":855,"comment_status":"","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"sns_share_botton_hide":"","vkExUnit_sns_title":"","_vk_print_noindex":"","sitemap_hide":"","_veu_custom_css":"","veu_display_promotion_alert":"","vkexunit_cta_each_option":"","_lightning_design_setting":[],"footnotes":""},"categories":[3],"tags":[8,17,16,319,56,26,15,685],"class_list":["post-1248","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-columns","tag-ai","tag-llm","tag-56","tag-26"],"veu_head_title_object":{"title":"","add_site_title":""},"_links":{"self":[{"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/posts\/1248","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/comments?post=1248"}],"version-history":[{"count":0,"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/posts\/1248\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/media\/855"}],"wp:attachment":[{"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/media?parent=1248"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/categories?post=1248"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/tags?post=1248"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}