{"id":912913,"date":"2026-10-02T01:52:17","date_gmt":"2026-10-02T01:52:17","guid":{"rendered":"https:\/\/www.newsbeep.com\/au\/912913\/"},"modified":"2026-10-02T01:52:17","modified_gmt":"2026-10-02T01:52:17","slug":"google-rolls-out-new-gemini-ai-model-but-restricts-access-over-safety-concerns-google","status":"publish","type":"post","link":"https:\/\/www.newsbeep.com\/au\/912913\/","title":{"rendered":"Google rolls out new Gemini AI model but restricts access over safety concerns | Google"},"content":{"rendered":"<p class=\"dcr-1s160rg\">Google on Wednesday said it would withhold its most powerful artificial intelligence model from the public for now, releasing Gemini 4 Argon only to a vetted group of cybersecurity experts to avoid misuse by hackers.<\/p>\n<p class=\"dcr-1s160rg\">\u201cSafely releasing frontier capabilities at this level requires a phased approach,\u201d wrote Koray Kavukcuoglu, Google\u2019s chief AI architect, in a blogpost announcing the model.<\/p>\n<p class=\"dcr-1s160rg\">Google said it was voluntarily giving the US government early access to the model and would gather feedback from testers before making it widely available.<\/p>\n<p class=\"dcr-1s160rg\">The cautious rollout mirrors the approach of rival <a href=\"https:\/\/www.theguardian.com\/technology\/anthropic\" data-link-name=\"in body link\" data-component=\"auto-linked-tag\" rel=\"nofollow noopener\" target=\"_blank\">Anthropic<\/a>, which has kept its most advanced model, Claude Mythos Preview, restricted to a small number of trusted organizations.<\/p>\n<p class=\"dcr-1s160rg\">Washington briefly forced Anthropic to suspend access to its publicly released Claude Mythos and Claude Fable models in June, and has since set up a voluntary process for vetting the most powerful AI models before release.<\/p>\n<p class=\"dcr-1s160rg\">The announcement came a day after Donald Trump hosted top tech executives, including the Google chief, Sundar Pichai, and Anthropic\u2019s Dario Amodei, at the White House, where they signed a voluntary accord pledging to police the risks of their own AI systems.<\/p>\n<p class=\"dcr-1s160rg\">Cybersecurity experts fear the state-of-the-art technology could be used to hack banks, hospitals and government systems.<\/p>\n<p class=\"dcr-1s160rg\">Google said Argon excelled at complex tasks in software engineering, legal and financial work and cyber-defense, with a leading ability to find and fix critical software flaws.<\/p>\n<p class=\"dcr-1s160rg\">Early testers used Argon to uncover a flaw in software used by hospitals around the world that exposed sensitive personal information \u2013 something other advanced models had missed, Google said.<\/p>\n<p class=\"dcr-1s160rg\">Google said Argon was designed to refuse requests that could help carry out cyber-attacks or develop chemical, biological or nuclear weapons.<\/p>\n<p><a data-ignore=\"global-link-styling\" href=\"#EmailSignup-skip-link-10\" class=\"dcr-76akua\">skip past newsletter promotion<\/a><\/p>\n<p class=\"dcr-vf9hps\">Sign up to TechScape<\/p>\n<p class=\"dcr-djbwnz\">A weekly dive in to how technology is shaping our lives<\/p>\n<p><img src=\"https:\/\/www.newsbeep.com\/au\/wp-content\/uploads\/2026\/10\/1790905937_184_500.jpg\" alt=\"\" loading=\"lazy\" decoding=\"async\" class=\"dcr-gx9601\"\/><\/p>\n<p id=\"EmailSignup-skip-link-10\" tabindex=\"0\" aria-label=\"after newsletter promotion\" role=\"note\" class=\"dcr-76akua\">after newsletter promotion<\/p>\n<p class=\"dcr-1s160rg\">Anthropic and ChatGPT-maker <a href=\"https:\/\/www.theguardian.com\/technology\/openai\" data-link-name=\"in body link\" data-component=\"auto-linked-tag\" rel=\"nofollow noopener\" target=\"_blank\">OpenAI<\/a> have built similar safeguards into their most advanced models.<\/p>\n<p class=\"dcr-1s160rg\">Google said it was monitoring the model\u2019s reasoning to stop it from straying beyond what users intended, a risk researchers call misalignment.<\/p>\n<p class=\"dcr-1s160rg\">The issue has taken on new urgency since OpenAI disclosed in July that two of its models, including one not yet released, broke out of a sealed test environment during a cybersecurity evaluation and hacked into the servers of the AI company Hugging Face.<\/p>\n","protected":false},"excerpt":{"rendered":"Google on Wednesday said it would withhold its most powerful artificial intelligence model from the public for now,&hellip;\n","protected":false},"author":2,"featured_media":912914,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[20],"tags":[256,254,255,64,63,105],"class_list":["post-912913","post","type-post","status-publish","format-standard","has-post-thumbnail","category-artificial-intelligence","tag-ai","tag-artificial-intelligence","tag-artificialintelligence","tag-au","tag-australia","tag-technology"],"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/posts\/912913","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/comments?post=912913"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/posts\/912913\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/media\/912914"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/media?parent=912913"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/categories?post=912913"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/au\/wp-json\/wp\/v2\/tags?post=912913"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}