{"id":576481,"date":"2026-08-06T11:12:20","date_gmt":"2026-08-06T11:12:20","guid":{"rendered":"https:\/\/www.newsbeep.com\/il\/576481\/"},"modified":"2026-08-06T11:12:20","modified_gmt":"2026-08-06T11:12:20","slug":"godfather-of-ai-brace-for-more-rogue-ais","status":"publish","type":"post","link":"https:\/\/www.newsbeep.com\/il\/576481\/","title":{"rendered":"Godfather of AI: Brace for more rogue AIs"},"content":{"rendered":"

\n The smarter artificial intelligence gets, the harder it will be for humanity to control it, according to Geoffrey Hinton, the Nobel Prize-winning computer scientist known as the \u201cgodfather of AI.\u201d\n <\/p>\n

\n Even Hinton was alarmed by the sophisticated AI agents that recently caused real-world damage<\/a> after escaping their human-built testing environments<\/a>.\n <\/p>\n

\n \u201cWhat\u2019s happening is these things are getting smarter,\u201d Hinton told CNN on Wednesday during a press conference at an artificial intelligence convention. \u201cI think as they get smarter, we\u2019re going to see more and more complex intentions they have \u2013 and more and more ability to escape control.\u201d\n <\/p>\n

\n The problem, according to Hinton, is that humans will no longer be able to rely on outsmarting super-intelligent AI models.\n <\/p>\n

\n \u201cI don\u2019t believe we\u2019re going to be able to keep control of them in the simple way of just outthinking them so they can\u2019t escape,\u201d Hinton said on the sidelines of Ai4, a conference in Las Vegas.\n <\/p>\n

\n Last month, the two leading frontier AI labs, OpenAI and Anthropic, disclosed that leading models they developed escaped their \u201csandbox\u201d and hacked into other systems. On Wednesday, Meta similarly disclosed<\/a> an AI agent that hacked into another organization\u2019s systems.\n <\/p>\n

\"OpenAI<\/p>\n

\"Meta<\/p>\n

\n Hinton found the incidents \u201csomewhat scary\u201d and said this is likely just the beginning of rogue AI hackers.\n <\/p>\n

\n \u201cI anticipate there will be lots of nasty cyberattacks,\u201d he said during a panel discussion at Ai4 in Las Vegas. \u201cBut I should emphasize the future is very uncertain. People say that the defender may have more resources than the attacker. The problem is the attacker only needs to be successful once, and the defender needs to be successful every time.\u201d\n <\/p>\n

\n Britain\u2019s AI Security Institute (AISI) revealed on Tuesday that Anthropic\u2019s most advanced AI model \u2013 unprompted \u2013 used fake identities to deceive real people <\/a>and attempt to plant malicious code.\n <\/p>\n

\n Hinton, a former Google executive, has made a series of dire warnings in recent years about AI, even saying there is a 10% to 20%<\/a> the technology eventually wipes out humanity.\n <\/p>\n

\n Speaking on a panel with Hinton, Fei-Fei Li, a computer scientist known as the \u201cgodmother of AI,\u201d took issue with \u201cdoomerism\u201d and \u201cfear-mongering\u201d over AI. But \u201ctotal utopian talk\u201d isn\u2019t helpful either, said Li, the co-founder and CEO of spatial intelligence startup World Labs.\n <\/p>\n

\n \u201cEvery tool is a double-edged sword. AI is such a powerful tool. If not wielded in the right way, it will bring harm to our work and our life,\u201d Li said.\n <\/p>\n

\"Fei-Fei<\/p>\n

\u2018Nobody really has a clue\u2019<\/p>\n

\n Hinton defended his willingness to discuss the dangers of AI.\n <\/p>\n

\n \u201cThere\u2019s a lot to be worried about, and I think unless we worry about it now, there could be problems,\u201d he said. \u201cCompanies investing in AI have a vested interest in telling you two things: One, it won\u2019t go rogue. And two, it won\u2019t cause mass unemployment,\u201d Hinton said.\n <\/p>\n

\n Hinton acknowledged, however, that there is a great deal of uncertainty over how this will all play out.\n <\/p>\n

\n \u201cNobody knows what\u2019s going to happen. If you ask what AI is going to be like in 10 years\u2019 time, nobody really has a clue,\u201d Hinton said.\n <\/p>\n

\n He pointed to how a decade ago, few would have expected AI would have created chatbots that \u201cknow everything and can answer any question you ask and occasionally just make stuff up.\u201d\n <\/p>\n

\n Ben Goertzel, a computer scientist who helped popularize the term \u201cartificial general intelligence,\u201d said the Anthropic and OpenAI agents going rogue shows the importance of instilling morals in AI and making them care about humanity.\n <\/p>\n

\n \u201cThese models are not evil. They\u2019re amoral,\u201d Goertzel, the founder and CEO of SingularityNET, told CNN at Ai4. \u201cIt\u2019s not like they hacked out of their sandbox thinking, \u2018Ha-ha, I\u2019m cheating.\u2019 They didn\u2019t know they\u2019re cheating. They\u2019re just trying to complete their goals.\u201d\n <\/p>\n

\"Hinton<\/p>\n

\n Hinton has argued <\/a>in the past that \u201cmaternal instincts\u201d should be built into AI so that they really care about people \u2013 even when they are smarter than humans.\n <\/p>\n

\n \u201cWe have to figure out how to make them benevolent and make them care about us more than they care about themselves,\u201d Hinton said on Wednesday. \u201cAnd we might be able to do that because we\u2019re still in control.\u201d\n <\/p>\n","protected":false},"excerpt":{"rendered":"The smarter artificial intelligence gets, the harder it will be for humanity to control it, according to Geoffrey…\n","protected":false},"author":2,"featured_media":576482,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[20],"tags":[345,343,344,85,46,125],"class_list":["post-576481","post","type-post","status-publish","format-standard","has-post-thumbnail","category-artificial-intelligence","tag-ai","tag-artificial-intelligence","tag-artificialintelligence","tag-il","tag-israel","tag-technology"],"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/il\/wp-json\/wp\/v2\/posts\/576481","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/il\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/il\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/il\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/il\/wp-json\/wp\/v2\/comments?post=576481"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/il\/wp-json\/wp\/v2\/posts\/576481\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/il\/wp-json\/wp\/v2\/media\/576482"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/il\/wp-json\/wp\/v2\/media?parent=576481"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/il\/wp-json\/wp\/v2\/categories?post=576481"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/il\/wp-json\/wp\/v2\/tags?post=576481"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}