Calls for a global slowdown in artificial intelligence development have gained support from leading technology figures, but have so far met little substantive response in Washington.
Dario Amodei, chief executive of Anthropic, issued the appeal in a 3,800-word essay on Saturday, renewing concerns about the dangers posed by increasingly powerful AI systems. Elon Musk, a rival in the AI race who remains close to President Donald Trump, responded on X: “Dario is right.”
Sam Altman, chief executive of OpenAI, also expressed support for Mr Amodei’s position.
The intervention followed the highly public resignation of Jacob Coxon, an Anthropic engineer who had previously worked for OpenAI. In a social media post, Mr Coxon accused the two leading AI laboratories of expressing deep concern while continuing to accelerate towards “superhuman systems that can hack anything, revolutionize any field overnight, and acquire real power and resources.”
Despite the warnings from figures within the industry, the US government’s response has remained limited. The contrast has been drawn with the decades after the Second World War, when Washington and other governments sought to reduce the risk of nuclear conflict through new institutions and arms-control treaties.
Some computer scientists believe the warnings about AI are overstated. Mr Trump has taken a prominent position among those resisting such concerns, declining to engage with the question of how potential risks should be balanced against the technology’s possible benefits.
Asked on Thursday whether he had any concerns about AI leading to human extinction, the president replied: “No, I don’t have any.”
He instead pointed to what he regards as the overriding strategic danger: the United States losing the AI race to China.
“I have concerns that if we don’t win A.I., we’re going to be put in a very bad position,” Mr Trump said, before turning to the argument that any American slowdown could allow China to take the lead.
