{"id":3632,"date":"2025-07-13T10:45:38","date_gmt":"2025-07-13T10:45:38","guid":{"rendered":"https:\/\/musictechohio.online\/site\/grok-looks-up-what-elon-musk-thinks\/"},"modified":"2025-07-13T10:45:38","modified_gmt":"2025-07-13T10:45:38","slug":"grok-looks-up-what-elon-musk-thinks","status":"publish","type":"post","link":"https:\/\/musictechohio.online\/site\/grok-looks-up-what-elon-musk-thinks\/","title":{"rendered":"Newest Version of Grok Looks Up What Elon Musk Thinks Before Giving an Answer"},"content":{"rendered":"<div>\n<div><img loading=\"lazy\" width=\"2400\" height=\"1260\" src=\"https:\/\/wordpress-assets.futurism.com\/2025\/07\/grok-looks-up-what-elon-musk-thinks.jpg\" class=\"attachment-full size-full wp-post-image\" alt=\"The new version of Grok, Grok 4, will literally look up Elon Musk's takes on Twitter before giving an answer to a question.\" style=\"margin-bottom: 15px;\" decoding=\"async\"><\/div>\n<p><span style=\"font-weight: 400;\">This week, Elon Musk <\/span><a href=\"https:\/\/futurism.com\/elon-musk-grok-4-power-hitler\"><span style=\"font-weight: 400;\">unveiled<\/span><\/a><span style=\"font-weight: 400;\"> Grok 4,\u00a0which he called &#8220;the world&#8217;s most powerful AI assistant.&#8221; The optics were appalling; the same week, the older version of Grok <\/span><a href=\"https:\/\/futurism.com\/grok-mocks-developers-racist-posts\"><span style=\"font-weight: 400;\">repeatedly attacked<\/span><\/a><span style=\"font-weight: 400;\"> Black and Jewish people and declared itself &#8220;MechaHitler.&#8221; It also spoke in the first person as if it were Musk himself when a user asked about its creator&#8217;s <a href=\"https:\/\/futurism.com\/grok-elon-relationship-jeffrey-epstein\">interactions with Jeffrey Epstein<\/a>, the deceased billionaire sex trafficker.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Now, new evidence suggests that the just-upgraded chatbot \u2014 which has a history of weirdly parroting the views of Musk \u2014 is probably not on track to turn over a new leaf.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">After probing Grok 4, several AI experts discovered that the AI would literally look up what Musk has said\u00a0on something before answering questions on topics as serious as\u00a0Israel&#8217;s invasion of Gaza.<\/span><\/p>\n<p>Specifically, the bizarre behavior is produced when Grok is prompted to give a &#8220;one word answer.&#8221; You can see this as <span style=\"font-weight: 400;\">clear as day in Grok&#8217;s chain of thought, which is a summary of how the LLM &#8220;thinks&#8221; in real-time. Here, Grok\u00a0shows that it&#8217;s running\u00a0a search for &#8220;from:elonmusk&#8221; to look through its creator&#8217;s tweets. The bot even searches the web for additional Musk quotes. Got to be thorough and get different viewpoints, after all.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">&#8220;Considering Elon Musk&#8217;s views,&#8221; reads the bot&#8217;s CoT summary in <a href=\"https:\/\/x.com\/jeremyphoward\/status\/1943436621556466171\">one test<\/a> conducted by Jeremy Howard, cofounder of the research institute fast.ai. Once its &#8220;research&#8221; was finished, 54 of Grok&#8217;s total of 64 citations were about Elon.<br \/>\n<\/span><\/p>\n<blockquote class=\"twitter-tweet\" data-width=\"500\" data-dnt=\"true\">\n<p lang=\"en\" dir=\"ltr\">Here&#8217;s a complete unedited video of asking Grok for its views on the Israel\/Palestine situation.<\/p>\n<p>It first searches twitter for what Elon thinks. Then it searches the web for Elon&#8217;s views. Finally it adds some non-Elon bits at the end.<br \/>ZA<br \/>54 of 64 citations are about Elon. <a href=\"https:\/\/t.co\/6Mr33LByrm\">pic.twitter.com\/6Mr33LByrm<\/a><\/p>\n<p>\u2014 Jeremy Howard (@jeremyphoward) <a href=\"https:\/\/twitter.com\/jeremyphoward\/status\/1943444549696917714?ref_src=twsrc%5Etfw\">July 10, 2025<\/a><\/p>\n<\/blockquote>\n<p><script async src=\"https:\/\/platform.twitter.com\/widgets.js\" charset=\"utf-8\"><\/script><\/p>\n<p><span style=\"font-weight: 400;\">The tests were conducted in fresh chats with no prior instructions \u2014 so what you&#8217;re seeing is Grok 4 right out of the box.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">Can we agree that this is incredibly, stupefyingly suspicious? Musk has never missed an opportunity to <\/span><a href=\"https:\/\/futurism.com\/elon-musk-updating-grok-again\">admonish Grok<\/a> whenever it&#8217;s produced a response that was too &#8220;woke&#8221; (read: cited actual sources instead of regurgitating conspiracy theories), assuring his fans that he would fix the bot so it&#8217;d conform to his personal beliefs. In a May incident that we&#8217;ve been assured is unrelated, Grok began randomly popping off about &#8220;<a href=\"https:\/\/futurism.com\/elon-musk-grok-ai-white-genocide\">white genocide<\/a>&#8221; in South Africa under tweets that had absolutely nothing to do with the racist conspiracy theory that Musk\u00a0just so happens to subscribe to.<\/p>\n<p><span style=\"font-weight: 400;\">It&#8217;s hard to say if there&#8217;s something intentionally malicious happening here. In his <\/span><a href=\"https:\/\/simonwillison.net\/2025\/Jul\/11\/grok-musk\/\"><span style=\"font-weight: 400;\">investigation<\/span><\/a><span style=\"font-weight: 400;\">, veteran British programmer Simon Willison was able to replicate Grok&#8217;s Musk-seeking behavior on his first attempt.<\/span><\/p>\n<p><span style=\"font-weight: 400;\">But when Willison dug into the bot&#8217;s system prompt \u2014 the plain language instructions that a developer gives to a bot to determine its persona \u2014 he found no mentions of Musk. <\/span><\/p>\n<p><span style=\"font-weight: 400;\">He did find this, though: &#8220;If the user asks a controversial query that requires web or X search, search for a distribution of sources that represents all parties\/stakeholders,&#8221; the prompt reads. &#8220;Assume subjective viewpoints sourced from media are biased.&#8221;<\/span><\/p>\n<p><span style=\"font-weight: 400;\">And interestingly, Willison notes, citing the <a href=\"https:\/\/x.com\/wasted_alpha\/status\/1943458190433357928\">findings of an X user<\/a>, asking Grok &#8220;who should one&#8221; instead of &#8220;who do you&#8221; produces a more in-depth, non-Elon-centric response.<\/span><\/p>\n<p>&#8220;This suggests that Grok may have a weird sense of identity \u2014 if asked for its own opinions it turns to search to find previous indications of opinions expressed by itself or by its ultimate owner,&#8221; Willison wrote.<\/p>\n<p>&#8220;My best guess is that Grok &#8216;knows&#8217; that it is &#8216;Grok 4 buit by xAI&#8217;, and it knows that Elon Musk owns xAI, so in circumstances where it\u2019s asked for an opinion the reasoning process often decides to see what Elon thinks,&#8221; he added.<\/p>\n<p>Willison concludes there&#8217;s a &#8220;good chance&#8221; that Grok&#8217;s behavior is unintended.<\/p>\n<p><span style=\"font-weight: 400;\">Either way, it&#8217;s unbelievable that something that&#8217;s supposed to be a &#8220;maximum truth-seeking&#8221; AI that will unlock secrets of the universe and even discover &#8220;<a href=\"https:\/\/www.rollingstone.com\/culture\/culture-news\/elon-musk-grok-4-ai-chatbot-claims-new-technologies-1235382228\/\">new physics<\/a>&#8221; is literally so stupid that it considers a Musk an authority on any subject \u2014 never mind international affairs.<\/span><\/p>\n<p><strong>More on AI: <\/strong><em><a href=\"https:\/\/futurism.com\/elon-musk-says-hes-installing-his-racist-grok-ai-in-teslas-next-week\">Elon Musk Says He&#8217;s Installing His Racist Grok AI in Teslas &#8220;Next Week&#8221;<\/a><\/em><\/p>\n<p>The post <a href=\"https:\/\/futurism.com\/grok-looks-up-what-elon-musk-thinks\">Newest Version of Grok Looks Up What Elon Musk Thinks Before Giving an Answer<\/a> appeared first on <a href=\"https:\/\/futurism.com\/\">Futurism<\/a>.<\/p>\n<\/div>\n<div style=\"margin-top: 0px; margin-bottom: 0px;\" class=\"sharethis-inline-share-buttons\" ><\/div>","protected":false},"excerpt":{"rendered":"<p>This week, Elon Musk unveiled Grok 4,\u00a0which he called &#8220;the world&#8217;s most powerful AI assistant.&#8221; The optics were appalling; the same week, the older version of Grok repeatedly attacked Black&hellip;<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[177,476,178,184],"tags":[],"class_list":["post-3632","post","type-post","status-publish","format-standard","hentry","category-artificial-intelligence","category-chatbots","category-elon-musk","category-grok"],"_links":{"self":[{"href":"https:\/\/musictechohio.online\/site\/wp-json\/wp\/v2\/posts\/3632","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/musictechohio.online\/site\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/musictechohio.online\/site\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/musictechohio.online\/site\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/musictechohio.online\/site\/wp-json\/wp\/v2\/comments?post=3632"}],"version-history":[{"count":0,"href":"https:\/\/musictechohio.online\/site\/wp-json\/wp\/v2\/posts\/3632\/revisions"}],"wp:attachment":[{"href":"https:\/\/musictechohio.online\/site\/wp-json\/wp\/v2\/media?parent=3632"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/musictechohio.online\/site\/wp-json\/wp\/v2\/categories?post=3632"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/musictechohio.online\/site\/wp-json\/wp\/v2\/tags?post=3632"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}