The U.S. Treasury Secretary, Scott Bessent, criticized the warnings from prominent figures in artificial intelligence about the existential risks the technology poses, labeling them alarmist and useless. He called on the industry itself to self-regulate and present solutions.
“I believe we have to be prepared for any scenario, but, on the other hand, this alarmism without solutions from some members of the AI community is not leadership,” Bessent told Axios in an interview published on Saturday (3).
The CEO of Anthropic PBC, Dario Amodei, the head of OpenAI, Sam Altman, and Elon Musk of xAI Corp., called in September for the industry to slow down the development of its most advanced models, citing growing risks to humanity. Microsoft cofounder Bill Gates also warned that the technology is powerful enough to cause “a billion deaths” and called for government regulation.
The president Donald Trump rejected almost immediately such calls, labeling the increasingly somber warnings about the risk to humanity as a “hoax” and insisting that the U.S. needs to advance to maintain leadership over China — a sentiment echoed by Bessent in his Axios remarks. “We cannot lose our leadership to the Chinese,” he said.
“It’s the people in the labs who need to take responsibility, and I agree with that,” said Bessent. “I think the labs are also starting to think that way.”
No New Regulations
In a meeting with the industry’s top executives, including Musk, the CEO of Microsoft, Satya Nadella, Mark Zuckerberg, of Meta Platforms Inc., the CEO of Palantir Technologies Inc., Alex Karp, and the president of OpenAI, Greg Brockman, the American president dismissed the idea of new federal regulations.
“There is a belief that there should be huge self-regulation, and automatically we already have regulation with the Department of Justice, the FBI and all of that, but self-regulation is very important,” Trump told reporters on the occasion. Instead, he endorsed the use of external auditors to assess the safety of AI systems in a non-binding agreement with Silicon Valley leaders.
Bessent told Axios that the U.S. will also seek to mitigate risks with a bilateral notification system with China, aimed at reducing uncertainties, sharing alerts about serious incidents, and helping both governments respond before an AI problem worsens or spreads globally.
The Treasury Secretary and Chinese Vice Premier He Lifeng discussed AI during talks in New York in September, and said the two countries agreed to create an “U.S.-China Dialogue on AI” to try to develop a common understanding of the benefits and threats of the technology. The two met shortly before the summit between Presidents Donald Trump and Xi Jinping in Washington, where AI was among the topics.
Xi told Trump during his visit that China and the U.S. can continue discussing AI, exchanging views on its risks and benefits and working together to prevent its misuse, Xinhua reported.
Bessent told Axios that they discussed risks such as uncontrolled AI agents and non-state actors using AI for cyber or biological threats.
“Many of these Chinese models are very powerful — not as powerful as the American models — but you end up with something that has 80%, 90% of the power, with no protective barriers, that can be imported to any place in the world,” Bessent told the outlet.
Asked about calls for regulation by AI industry executives, Bessent answered: “Well, then they should slow down.”
“We want a safe acceleration,” he said. “It is the people in the laboratories who need to take responsibility, and I agree with that.”
© 2026 Bloomberg L.P.
