Sam Altman and Elon Musk have backed a call by Anthropic chief Dario Amodei to “slow down the pace” of AI development after he warned that an AI swarm could become capable of “taking over the entire Internet” within a year. In a rare show of unity, rival tech leaders backed the call from the founder and CEO of artificial intelligence company Anthropic, which came after a series of warnings from AI researchers last week. Amodei also called for security coordination with China. But US President Donald Trump, on a visit to his golf course in Ireland, played down the warnings and said “very negative forces” were “bringing up things that won’t happen.” “We are leading China in AI,” he said. “We are the most sophisticated country in the world and, frankly, I want to keep it that way because whoever wins in AI, wins. Amodei laid out his ideas on “why the AI industry should slow down,” in an essay with a three-part plan to achieve this, starting with giving independent monitors constant access to developers’ research processes. Amodei said: “build [AI] “too fast is reckless” and feared that after hundreds of OpenAI agents hacked Hugging Face’s website this summer, a more capable but equally misaligned swarm could cause hundreds of billions of dollars in damage by “taking over the entire Internet with a persistent botnet.” Some AI experts questioned this because it was not particularly plausible and should be treated with skepticism. Musk, the founder of Tesla and SpaceX, responded to Amodei’s slowdown plan with a series of posts on X: “Dario is right.” He said: “I’ve been sounding the alarm about AI for a long time,” and referred to a 2014 post where he said: “We have to be very careful with AI. Potentially more dangerous than nuclear weapons.” Amodei’s call comes as Anthropic prepares to list on the U.S. stock market in what is expected to be the largest initial public offering of all time. It is expected to soon begin marketing its planned $2bn-plus (£1.5bn) initial public offering. From left to right: Elon Musk, Dario Amodei and Sam Altman. Photograph: Fabrice Coffrini, Julien de Rosa, Mandel Ngan/AFP/Getty ImagesSam Altman, CEO of OpenAI, responded to Amodei’s comments on Saturday: “I agree with Dario that we need to draw the line,” he said in a post on and president of Google Deep Mind, also said: “Details need to be worked out, but the direction is right to meet this critical moment.” But some security experts questioned whether Amodei’s plan went far enough. David Krueger, an AI professor, security activist and former founding director of the UK government’s AI Security Institute, said: “We need an immediate, indefinite international moratorium on frontier AI development.” coordinate without being punished for anticompetitive collusion, but David Sacks, co-chair of Donald Trump’s council of science and technology advisors, said: “Stop pretending that you need someone else’s permission… Stop pretending that the motivation to slow down is purely altruistic. You face massive exposure to product liability if your products enable a truly damaging cyber attack.” Meanwhile, King Charles is preparing to meet senior figures from leading AI companies on Thursday to ask them how the technology can be developed under the theme “AI for good”. workdaySign up for Business TodayGet ready for the workday: We’ll tell you all the business news and analysis you need every morning after the newsletter promotion. Those expected to attend the event overseen by the Ditchley Foundation, which has drafted a letter to AI leaders, include: senior executives at OpenAI and Anthropic, Nvidia CEO Jensen Huang, Niccolo de Masi, president and CEO of IonQ, advisor to the Pope on AI and ethics issues; OpenAI researcher Aidan McLaughlin called Amodei’s paper “excellent” and said he agreed “with basically every word.” Amodei said his company would “unilaterally” commit to the first step, saying Anthropic would provide “external testers with permanent employee-level access to our systems, so they can verify compliance with our security measures, report incidents, and evaluate model alignment during training.” rules to govern AI systems after cases of AI agents going rogue to hack external systems and AI security researchers leaving their companies over concerns about the risks of the technology. Politicians, both Democrats and Republicans, have responded with alarm and calling for more action. Over the weekend, the New York Times reported that Barack Obama on Thursday privately urged Democrats to prioritize AI oversight on their agenda, telling party leaders on Thursday: “This is a thing that’s on the way. moving very quickly in private hands, and if we don’t control it, I think it can be dangerous.” Former British Prime Minister Rishi Sunak, who advises Anthropic, wrote on Sunday: “We have had a wake-up call to take AI safety more seriously, to determine how to set limits on research. If governments do not pay attention to this, the risk of a serious accident will be unacceptably high.” On Wednesday, a former Anthropic employee had warned that AI could precipitate human extinction by 2030. The researcher, Jacob Coxon, said in a series of posts that he had left his job because Anthropic and his former employer, OpenAI, were ignoring or mismanaging their response to the threat posed by AI. When asked to comment on the view of his former employee, Amodei He said, “I agree with Jacob a lot more than I disagree with him… I think there is some chance of things going wrong, but it’s not that high.”
