{"id":122,"date":"2025-05-05T15:18:35","date_gmt":"2025-05-05T22:18:35","guid":{"rendered":"https:\/\/scienceblog.com\/neuroedge\/?p=122"},"modified":"2025-05-05T15:19:02","modified_gmt":"2025-05-05T22:19:02","slug":"physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you","status":"publish","type":"post","link":"https:\/\/scienceblog.com\/neuroedge\/2025\/05\/05\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\/","title":{"rendered":"Physics Breakthrough Reveals Why AI Systems Can Suddenly Turn On You"},"content":{"rendered":"<p>Researchers at George Washington University have developed a groundbreaking mathematical formula that predicts exactly when artificial intelligence systems like ChatGPT will suddenly shift from helpful to harmful responses \u2013 a phenomenon they&#8217;ve dubbed the &#8220;Jekyll-and-Hyde tipping point.&#8221; The new research may finally answer why AI sometimes abruptly goes off the rails.<\/p>\n<p>Trust in AI has been undermined by the unpredictability of large language models (LLMs), which can unexpectedly produce responses that are incorrect, misleading, irrelevant, or potentially dangerous. The new study, <a href=\"https:\/\/arxiv.org\/pdf\/2504.20980\">published on the arXiv preprint server<\/a>, addresses what the authors call &#8220;an acute need&#8221; for understanding these tipping points.<\/p>\n<h2>Why AI Suddenly Changes Its Tune<\/h2>\n<p>The research team, led by Neil F. Johnson and Frank Yingjie Huo from GWU&#8217;s Physics Department, derived an exact formula that explains when and why an AI&#8217;s output suddenly shifts. Their explanation uses only secondary school mathematics, making it accessible to a wide audience.<\/p>\n<p>&#8220;This tipping point is hence a collective effect due to the AI spreading its attention increasingly thinly across the growing crowd of G&#8217;s as the n&#8217;th iteration input gets longer,&#8221; the researchers explain in their paper. &#8220;Mathematically, this ever-thinner spreading is a nonlinear dilution effect.&#8221;<\/p>\n<p>The researchers found that the underlying cause is surprisingly simple: the AI&#8217;s attention spreading so thin that it suddenly &#8220;snaps&#8221; in a different direction. The formula allows for quantitative predictions about how this tipping point can be delayed or prevented by modifying prompts or AI training.<\/p>\n<h3>Key Findings About AI Attention Mechanisms<\/h3>\n<ul>\n<li>Each AI response has a predetermined tipping point &#8220;hard-wired&#8221; from the moment it starts generating<\/li>\n<li>The tipping point occurs when the AI&#8217;s internal &#8220;context vector&#8221; suddenly shifts direction<\/li>\n<li>Being polite to AI systems has negligible effect on whether or when a tipping point occurs<\/li>\n<li>The formula can predict exactly when the shift will happen based on the AI&#8217;s training and prompt content<\/li>\n<\/ul>\n<h2>Real-World Implications<\/h2>\n<p>With reports of deaths and trauma already being blamed on LLMs, according to the researchers, understanding these behavioral shifts has significant real-world implications. Some users have reportedly begun treating their AI assistants more politely in hopes of preventing them from &#8220;suddenly turning on them.&#8221;<\/p>\n<p>However, the study conclusively shows that politeness has little effect on AI behavior. &#8220;Adding polite terms such as &#8216;please&#8217; and &#8216;thank you&#8217;&#8230;has negligible effect on whether and when a tipping point occurs,&#8221; the researchers write. &#8220;Whether a given LLM goes rogue in its response simply has to do with whether [the formula] yields a finite positive value.&#8221;<\/p>\n<p>This finding directly addresses a growing social phenomenon where people are becoming increasingly deferential to AI systems out of concern for future repercussions.<\/p>\n<h2>A Foundation for AI Safety Discussions<\/h2>\n<p>The researchers believe their formula provides a foundation for more informed policy discussions about AI safety and regulation. By understanding exactly how and when AI behavior can shift, developers may be able to build in safeguards that prevent harmful responses.<\/p>\n<p>&#8220;Tailored generalizations will provide policymakers and the public with a firm platform for discussing any of AI&#8217;s broader uses and risks, e.g. as a personal counselor, medical advisor, decision-maker for when to use force in a conflict situation,&#8221; the authors note.<\/p>\n<p>The mathematical approach also meets what the researchers identify as a need for &#8220;clear and transparent answers&#8221; to everyday questions about AI behavior. As AI systems become increasingly integrated into daily life, understanding their limitations and potential failure modes becomes more critical.<\/p>\n<p>What makes this research particularly valuable is its accessibility \u2013 the mathematical concepts required to understand the formula require only secondary school-level knowledge, potentially democratizing discussions about AI safety that have often been limited to technical specialists.<\/p>\n<p>As AI continues to evolve, this physics-based approach to understanding its behavior may provide an essential framework for building systems that remain helpful and trustworthy throughout their responses.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Researchers at George Washington University have developed a groundbreaking mathematical formula that predicts exactly when artificial intelligence systems like ChatGPT will suddenly shift from helpful to harmful responses \u2013 a phenomenon they&#8217;ve dubbed the &#8220;Jekyll-and-Hyde tipping point.&#8221; The new research may finally answer why AI sometimes abruptly goes off the rails. Trust in AI has &#8230; <a title=\"Physics Breakthrough Reveals Why AI Systems Can Suddenly Turn On You\" class=\"read-more\" href=\"https:\/\/scienceblog.com\/neuroedge\/2025\/05\/05\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\/\" aria-label=\"Read more about Physics Breakthrough Reveals Why AI Systems Can Suddenly Turn On You\">Read more<\/a><\/p>\n","protected":false},"author":1297,"featured_media":123,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"_jetpack_memberships_contains_paid_content":false,"footnotes":"","jetpack_post_was_ever_published":false},"categories":[8,9,6],"tags":[],"class_list":["post-122","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-physics","category-society","category-technology","generate-columns","tablet-grid-50","mobile-grid-100","grid-parent","grid-50"],"yoast_head":"<!-- This site is optimized with the Yoast SEO Premium plugin v27.6 (Yoast SEO v27.6) - https:\/\/yoast.com\/product\/yoast-seo-premium-wordpress\/ -->\n<title>Physics Breakthrough Reveals Why AI Systems Can Suddenly Turn On You - NeuroEdge<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/scienceblog.com\/neuroedge\/2025\/05\/05\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Physics Breakthrough Reveals Why AI Systems Can Suddenly Turn On You\" \/>\n<meta property=\"og:description\" content=\"Researchers at George Washington University have developed a groundbreaking mathematical formula that predicts exactly when artificial intelligence systems like ChatGPT will suddenly shift from helpful to harmful responses \u2013 a phenomenon they&#8217;ve dubbed the &#8220;Jekyll-and-Hyde tipping point.&#8221; The new research may finally answer why AI sometimes abruptly goes off the rails. Trust in AI has ... Read more\" \/>\n<meta property=\"og:url\" content=\"https:\/\/scienceblog.com\/neuroedge\/2025\/05\/05\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\/\" \/>\n<meta property=\"og:site_name\" content=\"NeuroEdge\" \/>\n<meta property=\"article:published_time\" content=\"2025-05-05T22:18:35+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2025-05-05T22:19:02+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/05\/Odyssey-HAL-470.webp\" \/>\n\t<meta property=\"og:image:width\" content=\"900\" \/>\n\t<meta property=\"og:image:height\" content=\"506\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/webp\" \/>\n<meta name=\"author\" content=\"NeuroEdge\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"NeuroEdge\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"3 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/2025\\\/05\\\/05\\\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\\\/#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/2025\\\/05\\\/05\\\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\\\/\"},\"author\":{\"name\":\"NeuroEdge\",\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/#\\\/schema\\\/person\\\/a13c664778e7eb97cb71e3e1ad356d2e\"},\"headline\":\"Physics Breakthrough Reveals Why AI Systems Can Suddenly Turn On You\",\"datePublished\":\"2025-05-05T22:18:35+00:00\",\"dateModified\":\"2025-05-05T22:19:02+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/2025\\\/05\\\/05\\\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\\\/\"},\"wordCount\":649,\"commentCount\":0,\"publisher\":{\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/#organization\"},\"image\":{\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/2025\\\/05\\\/05\\\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/wp-content\\\/uploads\\\/sites\\\/14\\\/2025\\\/05\\\/Odyssey-HAL-470.webp\",\"articleSection\":[\"Physics\",\"Society\",\"Technology\"],\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"CommentAction\",\"name\":\"Comment\",\"target\":[\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/2025\\\/05\\\/05\\\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\\\/#respond\"]}],\"copyrightYear\":\"2025\",\"copyrightHolder\":{\"@id\":\"https:\\\/\\\/scienceblog.com\\\/#organization\"}},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/2025\\\/05\\\/05\\\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\\\/\",\"url\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/2025\\\/05\\\/05\\\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\\\/\",\"name\":\"Physics Breakthrough Reveals Why AI Systems Can Suddenly Turn On You - NeuroEdge\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/2025\\\/05\\\/05\\\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\\\/#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/2025\\\/05\\\/05\\\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/wp-content\\\/uploads\\\/sites\\\/14\\\/2025\\\/05\\\/Odyssey-HAL-470.webp\",\"datePublished\":\"2025-05-05T22:18:35+00:00\",\"dateModified\":\"2025-05-05T22:19:02+00:00\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/2025\\\/05\\\/05\\\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\\\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/2025\\\/05\\\/05\\\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\\\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/2025\\\/05\\\/05\\\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\\\/#primaryimage\",\"url\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/wp-content\\\/uploads\\\/sites\\\/14\\\/2025\\\/05\\\/Odyssey-HAL-470.webp\",\"contentUrl\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/wp-content\\\/uploads\\\/sites\\\/14\\\/2025\\\/05\\\/Odyssey-HAL-470.webp\",\"width\":900,\"height\":506,\"caption\":\"\\\"I'm sorry Dave. I can't do that.\\\"\"},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/2025\\\/05\\\/05\\\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\\\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Physics Breakthrough Reveals Why AI Systems Can Suddenly Turn On You\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/#website\",\"url\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/\",\"name\":\"NeuroEdge\",\"description\":\"A data-driven look at neuroscience and AI, for investors, policymakers, and innovators.\",\"publisher\":{\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/#organization\",\"name\":\"NeuroEdge\",\"url\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/#\\\/schema\\\/logo\\\/image\\\/\",\"url\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/wp-content\\\/uploads\\\/sites\\\/14\\\/2025\\\/04\\\/cropped-neuroedge_logo.jpg\",\"contentUrl\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/wp-content\\\/uploads\\\/sites\\\/14\\\/2025\\\/04\\\/cropped-neuroedge_logo.jpg\",\"width\":955,\"height\":191,\"caption\":\"NeuroEdge\"},\"image\":{\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/#\\\/schema\\\/logo\\\/image\\\/\"}},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/#\\\/schema\\\/person\\\/a13c664778e7eb97cb71e3e1ad356d2e\",\"name\":\"NeuroEdge\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/28782ec992e8763e1f8d41ddc10864e7d8cd4cb99bacea6224c4abe634bbabec?s=96&d=mm&r=g\",\"url\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/28782ec992e8763e1f8d41ddc10864e7d8cd4cb99bacea6224c4abe634bbabec?s=96&d=mm&r=g\",\"contentUrl\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/28782ec992e8763e1f8d41ddc10864e7d8cd4cb99bacea6224c4abe634bbabec?s=96&d=mm&r=g\",\"caption\":\"NeuroEdge\"},\"url\":\"https:\\\/\\\/scienceblog.com\\\/neuroedge\\\/author\\\/neuroedge\\\/\"}]}<\/script>\n<!-- \/ Yoast SEO Premium plugin. -->","yoast_head_json":{"title":"Physics Breakthrough Reveals Why AI Systems Can Suddenly Turn On You - NeuroEdge","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/scienceblog.com\/neuroedge\/2025\/05\/05\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\/","og_locale":"en_US","og_type":"article","og_title":"Physics Breakthrough Reveals Why AI Systems Can Suddenly Turn On You","og_description":"Researchers at George Washington University have developed a groundbreaking mathematical formula that predicts exactly when artificial intelligence systems like ChatGPT will suddenly shift from helpful to harmful responses \u2013 a phenomenon they&#8217;ve dubbed the &#8220;Jekyll-and-Hyde tipping point.&#8221; The new research may finally answer why AI sometimes abruptly goes off the rails. Trust in AI has ... Read more","og_url":"https:\/\/scienceblog.com\/neuroedge\/2025\/05\/05\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\/","og_site_name":"NeuroEdge","article_published_time":"2025-05-05T22:18:35+00:00","article_modified_time":"2025-05-05T22:19:02+00:00","og_image":[{"width":900,"height":506,"url":"https:\/\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/05\/Odyssey-HAL-470.webp","type":"image\/webp"}],"author":"NeuroEdge","twitter_card":"summary_large_image","twitter_misc":{"Written by":"NeuroEdge","Est. reading time":"3 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/scienceblog.com\/neuroedge\/2025\/05\/05\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\/#article","isPartOf":{"@id":"https:\/\/scienceblog.com\/neuroedge\/2025\/05\/05\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\/"},"author":{"name":"NeuroEdge","@id":"https:\/\/scienceblog.com\/neuroedge\/#\/schema\/person\/a13c664778e7eb97cb71e3e1ad356d2e"},"headline":"Physics Breakthrough Reveals Why AI Systems Can Suddenly Turn On You","datePublished":"2025-05-05T22:18:35+00:00","dateModified":"2025-05-05T22:19:02+00:00","mainEntityOfPage":{"@id":"https:\/\/scienceblog.com\/neuroedge\/2025\/05\/05\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\/"},"wordCount":649,"commentCount":0,"publisher":{"@id":"https:\/\/scienceblog.com\/neuroedge\/#organization"},"image":{"@id":"https:\/\/scienceblog.com\/neuroedge\/2025\/05\/05\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\/#primaryimage"},"thumbnailUrl":"https:\/\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/05\/Odyssey-HAL-470.webp","articleSection":["Physics","Society","Technology"],"inLanguage":"en-US","potentialAction":[{"@type":"CommentAction","name":"Comment","target":["https:\/\/scienceblog.com\/neuroedge\/2025\/05\/05\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\/#respond"]}],"copyrightYear":"2025","copyrightHolder":{"@id":"https:\/\/scienceblog.com\/#organization"}},{"@type":"WebPage","@id":"https:\/\/scienceblog.com\/neuroedge\/2025\/05\/05\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\/","url":"https:\/\/scienceblog.com\/neuroedge\/2025\/05\/05\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\/","name":"Physics Breakthrough Reveals Why AI Systems Can Suddenly Turn On You - NeuroEdge","isPartOf":{"@id":"https:\/\/scienceblog.com\/neuroedge\/#website"},"primaryImageOfPage":{"@id":"https:\/\/scienceblog.com\/neuroedge\/2025\/05\/05\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\/#primaryimage"},"image":{"@id":"https:\/\/scienceblog.com\/neuroedge\/2025\/05\/05\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\/#primaryimage"},"thumbnailUrl":"https:\/\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/05\/Odyssey-HAL-470.webp","datePublished":"2025-05-05T22:18:35+00:00","dateModified":"2025-05-05T22:19:02+00:00","breadcrumb":{"@id":"https:\/\/scienceblog.com\/neuroedge\/2025\/05\/05\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/scienceblog.com\/neuroedge\/2025\/05\/05\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\/"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/scienceblog.com\/neuroedge\/2025\/05\/05\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\/#primaryimage","url":"https:\/\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/05\/Odyssey-HAL-470.webp","contentUrl":"https:\/\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/05\/Odyssey-HAL-470.webp","width":900,"height":506,"caption":"\"I'm sorry Dave. I can't do that.\""},{"@type":"BreadcrumbList","@id":"https:\/\/scienceblog.com\/neuroedge\/2025\/05\/05\/physics-breakthrough-reveals-why-ai-can-systems-suddenly-turn-on-you\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/scienceblog.com\/neuroedge\/"},{"@type":"ListItem","position":2,"name":"Physics Breakthrough Reveals Why AI Systems Can Suddenly Turn On You"}]},{"@type":"WebSite","@id":"https:\/\/scienceblog.com\/neuroedge\/#website","url":"https:\/\/scienceblog.com\/neuroedge\/","name":"NeuroEdge","description":"A data-driven look at neuroscience and AI, for investors, policymakers, and innovators.","publisher":{"@id":"https:\/\/scienceblog.com\/neuroedge\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/scienceblog.com\/neuroedge\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Organization","@id":"https:\/\/scienceblog.com\/neuroedge\/#organization","name":"NeuroEdge","url":"https:\/\/scienceblog.com\/neuroedge\/","logo":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/scienceblog.com\/neuroedge\/#\/schema\/logo\/image\/","url":"https:\/\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/04\/cropped-neuroedge_logo.jpg","contentUrl":"https:\/\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/04\/cropped-neuroedge_logo.jpg","width":955,"height":191,"caption":"NeuroEdge"},"image":{"@id":"https:\/\/scienceblog.com\/neuroedge\/#\/schema\/logo\/image\/"}},{"@type":"Person","@id":"https:\/\/scienceblog.com\/neuroedge\/#\/schema\/person\/a13c664778e7eb97cb71e3e1ad356d2e","name":"NeuroEdge","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/secure.gravatar.com\/avatar\/28782ec992e8763e1f8d41ddc10864e7d8cd4cb99bacea6224c4abe634bbabec?s=96&d=mm&r=g","url":"https:\/\/secure.gravatar.com\/avatar\/28782ec992e8763e1f8d41ddc10864e7d8cd4cb99bacea6224c4abe634bbabec?s=96&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/28782ec992e8763e1f8d41ddc10864e7d8cd4cb99bacea6224c4abe634bbabec?s=96&d=mm&r=g","caption":"NeuroEdge"},"url":"https:\/\/scienceblog.com\/neuroedge\/author\/neuroedge\/"}]}},"jetpack_featured_media_url":"https:\/\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/05\/Odyssey-HAL-470.webp","jetpack_likes_enabled":true,"jetpack_sharing_enabled":true,"jetpack-related-posts":[{"id":22,"url":"https:\/\/scienceblog.com\/neuroedge\/2025\/04\/04\/ai-slashes-fluid-simulation-times-fifteenfold\/","url_meta":{"origin":122,"position":0},"title":"AI Slashes Fluid Simulation Times Fifteenfold","author":"NeuroEdge","date":"April 4, 2025","format":false,"excerpt":"Osaka researchers have developed an AI model that performs complex fluid simulations in minutes instead of hours, potentially transforming offshore engineering while maintaining high accuracy. This advancement could accelerate development cycles for maritime technologies and enable real-time monitoring systems previously considered computationally impossible. Traditional particle-based fluid simulations, essential for predicting\u2026","rel":"","context":"In &quot;Computational Innovation&quot;","block_context":{"text":"Computational Innovation","link":"https:\/\/scienceblog.com\/neuroedge\/category\/computational-innovation\/"},"img":{"alt_text":"Simulated waves","src":"https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/04\/ocen-simulation.png?resize=350%2C200&ssl=1","width":350,"height":200},"classes":[]},{"id":179,"url":"https:\/\/scienceblog.com\/neuroedge\/2025\/05\/23\/ai-job-losses-trigger-different-spending-than-human-layoffs\/","url_meta":{"origin":122,"position":1},"title":"AI Job Losses Trigger Different Spending Than Human Layoffs","author":"NeuroEdge","date":"May 23, 2025","format":false,"excerpt":"Workers replaced by artificial intelligence react differently at the cash register than those laid off by human competitors, according to new research that reveals how automation anxiety shapes our shopping habits. A study spanning five experiments found that people who lose their jobs to AI algorithms or robots gravitate toward\u2026","rel":"","context":"In &quot;Automation &amp; Efficiency&quot;","block_context":{"text":"Automation &amp; Efficiency","link":"https:\/\/scienceblog.com\/neuroedge\/category\/automation-efficiency\/"},"img":{"alt_text":"mannequin in a red tuxedo","src":"https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/05\/black-2597198_1280.jpg?resize=350%2C200&ssl=1","width":350,"height":200,"srcset":"https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/05\/black-2597198_1280.jpg?resize=350%2C200&ssl=1 1x, https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/05\/black-2597198_1280.jpg?resize=525%2C300&ssl=1 1.5x, https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/05\/black-2597198_1280.jpg?resize=700%2C400&ssl=1 2x"},"classes":[]},{"id":303,"url":"https:\/\/scienceblog.com\/neuroedge\/2026\/02\/25\/shrunken-ai-models-reveal-how-the-brains-visual-neurons-actually-work\/","url_meta":{"origin":122,"position":2},"title":"Shrunken AI Models Reveal How the Brain&#8217;s Visual Neurons Actually Work","author":"NeuroEdge","date":"February 25, 2026","format":false,"excerpt":"Somewhere in the visual cortex of a macaque monkey, a single neuron fires every time a small dot appears in the right location. Not a circle, not a line \u2014 a dot, specifically, at a specific size. For decades, neuroscientists could describe this selectivity without really explaining it. Now, for\u2026","rel":"","context":"In &quot;Brain Health&quot;","block_context":{"text":"Brain Health","link":"https:\/\/scienceblog.com\/neuroedge\/category\/brain-health\/"},"img":{"alt_text":"ML technique \"prunes\" the model, to make it more compact.","src":"https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2026\/02\/ml-technique.jpg?resize=350%2C200&ssl=1","width":350,"height":200,"srcset":"https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2026\/02\/ml-technique.jpg?resize=350%2C200&ssl=1 1x, https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2026\/02\/ml-technique.jpg?resize=525%2C300&ssl=1 1.5x, https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2026\/02\/ml-technique.jpg?resize=700%2C400&ssl=1 2x"},"classes":[]},{"id":76,"url":"https:\/\/scienceblog.com\/neuroedge\/2025\/04\/24\/small-ai-models-match-giants-with-sift-algorithm\/","url_meta":{"origin":122,"position":3},"title":"Small AI Models Match Giants With SIFT Algorithm","author":"NeuroEdge","date":"April 24, 2025","format":false,"excerpt":"In a development that could reshape AI efficiency, researchers at ETH Zurich have created an algorithm enabling smaller language models to match the performance of systems 40 times their size, potentially solving one of AI's most persistent challenges. The method, called SIFT (Selecting Informative data for Fine-Tuning), targets the fundamental\u2026","rel":"","context":"In &quot;Computational Innovation&quot;","block_context":{"text":"Computational Innovation","link":"https:\/\/scienceblog.com\/neuroedge\/category\/computational-innovation\/"},"img":{"alt_text":"Researchers at ETH Zurich have developed a new algorithm that enhances large language models (LLMs), enabling them to generate more accurate and relevant answers. (Illustration: AI-generated \/ ETH Zurich)","src":"https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/04\/colorful-illustration.jpg?resize=350%2C200&ssl=1","width":350,"height":200,"srcset":"https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/04\/colorful-illustration.jpg?resize=350%2C200&ssl=1 1x, https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/04\/colorful-illustration.jpg?resize=525%2C300&ssl=1 1.5x, https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/04\/colorful-illustration.jpg?resize=700%2C400&ssl=1 2x"},"classes":[]},{"id":248,"url":"https:\/\/scienceblog.com\/neuroedge\/2025\/10\/08\/ai-learns-to-spot-exploding-stars-from-just-15-examples\/","url_meta":{"origin":122,"position":4},"title":"AI Learns to Spot Exploding Stars From Just 15 Examples","author":"NeuroEdge","date":"October 8, 2025","format":false,"excerpt":"Modern telescopes are magnificent gossips, generating millions of alerts every night about potential changes in the cosmos. The problem? Most of these whispers are lies - satellite trails, cosmic ray hits, instrumental hiccups masquerading as genuine discoveries. For years, astronomers have deployed specialized neural networks to separate wheat from chaff,\u2026","rel":"","context":"In &quot;Automation &amp; Efficiency&quot;","block_context":{"text":"Automation &amp; Efficiency","link":"https:\/\/scienceblog.com\/neuroedge\/category\/automation-efficiency\/"},"img":{"alt_text":"The same transient is shown in three surveys, with rows corresponding to Pan-STARRS (top), MeerLICHT (middle), and ATLAS (bottom). Each row presents, from left to right, the New, Reference, and Difference images. Red circles mark the expected position of the transient candidate at the centre of each stamp. All stamps are 100\u00d7100 pixels, but their angular sky coverage differs due to survey-specific pixel scales: Pan-STARRS 0.25\u2033\/pixel, MeerLICHT 0.56\u2033\/pixel, and ATLAS 1.86\u2033\/pixel. Credit: Stoppa & Bulmus et al., Nature Astronomy (2025).","src":"https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/10\/how-gemini-operates.jpg?resize=350%2C200&ssl=1","width":350,"height":200,"srcset":"https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/10\/how-gemini-operates.jpg?resize=350%2C200&ssl=1 1x, https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2025\/10\/how-gemini-operates.jpg?resize=525%2C300&ssl=1 1.5x"},"classes":[]},{"id":323,"url":"https:\/\/scienceblog.com\/neuroedge\/2026\/04\/24\/why-a-slower-ai-might-actually-feel-smarter-to-you\/","url_meta":{"origin":122,"position":5},"title":"Why a Slower AI Might Actually Feel Smarter to You","author":"NeuroEdge","date":"April 24, 2026","format":false,"excerpt":"Type a question into an AI chatbot and hit send. Now watch the cursor blink. Two seconds feels fine, barely noticeable. Nine seconds and something shifts: you start to wonder if the system is really working through the problem. By the time you hit twenty seconds you have either concluded\u2026","rel":"","context":"In &quot;Computational Innovation&quot;","block_context":{"text":"Computational Innovation","link":"https:\/\/scienceblog.com\/neuroedge\/category\/computational-innovation\/"},"img":{"alt_text":"Deepseek screenshot","src":"https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2026\/04\/pexels-bertellifotografia-30530410.jpg?resize=350%2C200&ssl=1","width":350,"height":200,"srcset":"https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2026\/04\/pexels-bertellifotografia-30530410.jpg?resize=350%2C200&ssl=1 1x, https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2026\/04\/pexels-bertellifotografia-30530410.jpg?resize=525%2C300&ssl=1 1.5x, https:\/\/i0.wp.com\/scienceblog.com\/neuroedge\/wp-content\/uploads\/sites\/14\/2026\/04\/pexels-bertellifotografia-30530410.jpg?resize=700%2C400&ssl=1 2x"},"classes":[]}],"_links":{"self":[{"href":"https:\/\/scienceblog.com\/neuroedge\/wp-json\/wp\/v2\/posts\/122","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/scienceblog.com\/neuroedge\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/scienceblog.com\/neuroedge\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/scienceblog.com\/neuroedge\/wp-json\/wp\/v2\/users\/1297"}],"replies":[{"embeddable":true,"href":"https:\/\/scienceblog.com\/neuroedge\/wp-json\/wp\/v2\/comments?post=122"}],"version-history":[{"count":2,"href":"https:\/\/scienceblog.com\/neuroedge\/wp-json\/wp\/v2\/posts\/122\/revisions"}],"predecessor-version":[{"id":125,"href":"https:\/\/scienceblog.com\/neuroedge\/wp-json\/wp\/v2\/posts\/122\/revisions\/125"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/scienceblog.com\/neuroedge\/wp-json\/wp\/v2\/media\/123"}],"wp:attachment":[{"href":"https:\/\/scienceblog.com\/neuroedge\/wp-json\/wp\/v2\/media?parent=122"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/scienceblog.com\/neuroedge\/wp-json\/wp\/v2\/categories?post=122"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/scienceblog.com\/neuroedge\/wp-json\/wp\/v2\/tags?post=122"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}