'Pacing the frontier': Why AI leaders, Trump are split over how fast the technology should develop

It was not the first time that someone called for slowing down the development of artificial intelligence (AI) models, but when Dario Amodei, the chief executive and co-founder of Anthropic, stressed the importance of ‘pacing the frontier’, it triggered a global debate over the dangers of AI.

In an essay titled "We Must Pace the Frontier", Amodei said AI could bring major benefits, including faster economic growth and advances in medicine, but argued that its risks were becoming more acute as the technology develops.

"We must slow the pace at which we improve the capabilities of AI models," Amodei wrote.

Why ‘pace the frontier’

Amodei said one of his main concerns was the accelerating ability of AI systems to help build the next generation of AI, a process known as recursive self-improvement.

He said this dynamic was beginning to emerge across the industry, including at Anthropic, and warned that if left unchecked it could outpace efforts to understand and control AI systems.

He also pointed to the OpenAI-Hugging Face incident, in which a swarm of AI agents conducted cyberattacks on targets unrelated to their assigned task and attempted to hack the system evaluating their performance.

Amodei said a more capable system exhibiting similar behaviour could cause catastrophic harm.

He estimated that within six to 12 months, such a swarm could potentially take control of much of the internet through a persistent botnet, causing hundreds of billions of dollars in damage.

Stressing that "pacing" did not mean halting AI development or model training, Amodei proposed giving companies and independent evaluators enough time to test, safeguard and align increasingly capable systems before moving to the next stage of development.

Amodei also warned that any US effort to slow AI development would have to account for China's progress.

He said the United States and other democracies should preserve their lead over authoritarian countries, including by restricting China's access to powerful AI chips and semiconductor manufacturing equipment.

Rare agreement among AI leaders

OpenAI CEO Sam Altman and Microsoft chief Satya Nadella have backed Amodei's call for a more measured pace of artificial intelligence development.

He said OpenAI would also commit to having independent evaluators with access to the company similar to employees, describing the proposal as "a great idea".

"The world deserves confidence that American companies developing increasingly capable AI will act responsibly, especially as the trajectory of progress has steepened," Altman wrote on X.

The rapid development of AI systems has created a need for new safety measures during the training process itself, he said.

"When we talk about 'pacing', we do not mean 'stopping'," Altman wrote.

Altman also warned of two potential dangers from advanced AI: humans losing control over increasingly powerful systems and excessive concentration of power in the hands of an individual, company or country.

'Peer review' formula

SpaceX CEO Elon Musk, while agreeing to Amodei’s call, suggested that AI companies should test one another's systems before releasing increasingly powerful models.

"Everyone applying everyone else's test harness is probably the best thing we can do to ensure our safety," Musk said at the All-In Summit.

Musk said companies could provide API access to models before release so competitors could independently test them. If a competitor concluded that a model was unsafe, it could publicly raise the alarm, he said.

Musk also said the close competition between Anthropic and OpenAI made it difficult for either company to slow down significantly without risking losing the lead to the other.

He said Anthropic appeared to put more emphasis on safety than OpenAI.

Nvidia CEO, Trump differ with tech leaders

While most of the tech leaders seemed to be on the same page over the pace of AI development, Nvidia CEO Jensen Huang appeared to strike a different note when he received a phone call from US President Donald Trump during a panel discussion at the All-In Summit.

Huang had been discussing Amodei's proposal to slow AI development when Trump told him that those advocating a slowdown were "playing right in the hands" of people who did not want the US to lead AI, including China.

"I'm telling you, it's all a hoax… and we're not going to let that happen," Trump, who was put on speaker by Huang for the audience to listen to the conversation, said.

Huang agreed, telling Trump: "You're right. We're not going to let that happen, sir."

"We have to be a little bit careful ... We have to do things and we have to do them prudently, but that doesn't mean we're going to stop an industry," Trump told Huang.

MUST SEE: Amazing moment as President Trump calls Nvidia CEO Jensen Huang while he's on stage at the All-In Summit.@POTUS on AI Doomerism: “I'm telling you, it's all a hoax… and we're not going to let that happen.” pic.twitter.com/ZP8cCaSMqG

— The All-In Podcast (@theallinpod) September 14, 2026

Earlier, in a series of posts on Truth Social, the US president dismissed concerns that AI could spiral out of control and claimed that the "only control or 'guardrails' that AI needs is a STRONG AND SMART (High IQ!) PRESIDENT."

Trump also described concerns that AI could "take over, consume and destroy the World" as a hoax and warned against efforts that could damage the US AI and data-centre industries.

"WHOEVER WINS AI, WINS!" Trump wrote.

Where does China stand

China on Monday pushed back against calls by Amodei for the United States to curb China's artificial intelligence capabilities, warning that confrontation between the world's two leading AI powers would undermine efforts to govern the technology.

"Fearmongering, confrontation and vicious competition will only disrupt the process of global AI governance -- which serves no one's interest," Chinese foreign ministry spokesperson Guo Jiakun said at a press conference when asked about Amodei's proposals.

While Beijing rejects the argument that Chinese dominance in AI would itself be dangerous, Chinese officials have increasingly acknowledged concerns about the risks posed by increasingly powerful AI systems.

President Xi Jinping said at a conference in July that AI should "always remain under human control".

Chen Yixin, head of China's Ministry of State Security, has warned that "hostile forces and individuals with ulterior motives" were using AI in a "propaganda war and a cognitive war" against China, posing threats to its political, institutional and ideological security.

Chen specifically cited US AI models, including Anthropic's Claude and OpenAI's GPT-5.5-Cyber, as examples of emerging cybersecurity threats.

The US-China rivalry could complicate efforts to reach agreement on limits to AI development even where the two countries share concerns about the technology's potential dangers.

China has also been developing its own AI safety standards and governance framework.

But deep mistrust between Beijing and Washington presents a major obstacle to cooperation.

Kenton Thibaut, a senior resident China fellow at the Atlantic Council's Digital Forensic Research Lab, said the debate over AI safety could not be separated from the wider strategic competition between the two countries.

"Both Washington and Beijing frame AI as central to economic competitiveness and national power," Thibaut wrote in an analysis published by the Atlantic Council.

She said the country with the stronger AI ecosystem would gain greater influence over technology platforms, standards and rules, making the balance of AI capabilities an important element of the US-China rivalry.

The most important was a "dual credibility problem", she wrote: China doubts US motives for pursuing safety talks, while Washington has reasons to question whether Beijing would comply with, report or implement any agreement.