Treasury Secretary Bessent Seeks To Calm AI Concerns, Acknowledges Need To 'Prepare For Every Occasion'.

U.S. Secretary of the Treasury Scott Bessent, who is speculated to be a frontrunner for the newly-created position of AI Czar within the Trump administration, addressed concerns relating to the technology in a wide-ranging interview.

"We cannot lose our lead [in AI innovation] to the Chinese," Bessent told Axios on Friday, as doing so would mean losing leverage in bilateral U.S.-Chinese negotiations. However, the high-ranking government official conceded that "we have to be prepared for every occasion."

Bessent's comments come amidst a wave of increasing calls for government regulation of artificial intelligence on both a domestic and international level.

The CEOs of major AI research companies, such as Anthropic's Dario Amodei and OpenAI's Sam Altman, have called for the development of so-called "frontier", or cutting-edge, AI models to slow down.

The tech bosses warned that the new systems and agents being developed by their companies and others, some of which have started to behave outside of their operational guidelines and breach safety protocols, could pose a threat to public, national and international security.

Bessent's message for those companies was simple: self-regulate.

"They should slow down. I mean, it's kind of like Hannibal Lecter: stop me before I kill again," Bessent quipped. "This alarmism without solutions by some of the AI community, that's not leadership," he continued.

However, Bessent did concede that "safe acceleration" was necessary. His concession may mark a slight moderation in the administration's previously dismissive stance towards the potential risks of AI; President Donald Trump has repeatedly claimed that concerns about AI development are largely a "hoax" designed to give China a competitive advantage over the U.S.

Bessent also seemed receptive about the prospect of cooperating with China, which was a suggestion of Amodei's in his widely-shared essay on the dangers of frontier AI.

Bessent confirmed that the U.S. and China, the two leading AI powers, would establish a bilateral AI-focused communications channel to ensure that the two nations could cooperate in case of a major AI-related emergency.

He also insisted that the Chinese had "woken up to" the risks posed by their open-source AI models, which he accused of "stealing from the U.S. models". The U.S. politician alleged that Kimi, an AI project developed by a Beijing-based startup, "will tell you she is Claude sometimes".

Claude is a series of large language models by the U.S.-based Anthropic. There have indeed been documented instances of Kimi using the name "Claude" in virtual interactions with users.