Bessent Says US and China Open Super Intelligence Dialogue, No Limits Pledged
Treasury Secretary Scott Bessent says Washington and Beijing have set up a new channel on super intelligence, while making no promise to slow AI development.

Treasury Secretary Scott Bessent said Tuesday that Washington and Beijing have created a fresh channel for talks on super intelligence, or SI, according to CNBC Technology. He described the mechanism as a way to shield both countries from hazards they share and to prevent misreading between the two systems.
Speaking in New York at a gala held by a US-China relations group, Bessent said the two sides would trade methods for checking cutting-edge models before launch. He also said the aim is to keep human judgment in charge of the technology, and he expects each government to notify the other about grave incidents tied to SI models.
Bessent did not promise to cap how far the technology is pushed. The two nations remain in a fierce contest for leadership in artificial intelligence, and his remarks left that competition untouched.
The Trump administration has mostly avoided widening limits on AI work, fearing it would hand an advantage to China. The president recently showed some willingness to accept safeguards, telling CBS that guardrails do not bother him, while warning that rules should not leave the US far behind a rival he said faces no such constraints.
Beijing's foreign ministry rejected appeals in September to slow frontier models, arguing that alarm and confrontation would only unsettle worldwide AI governance. Huiyao Wang, who founded the Center for China and Globalization, a think tank close to Beijing's positions, said last month that the two countries should face AI dangers and handle them jointly.
During Xi's September visit, the two governments agreed to start an AI dialogue with a communications link for easing serious incidents. The leaders are set to meet for a third time next week, with Bessent attending, around the Asia-Pacific Economic Cooperation Leaders' Meeting on Nov. 18 and 19.
Bessent said both nations must show they can balance rivalry with the duties of leadership, and he noted that their interests and methods will not always align. China's ambassador to the US, Xie Feng, told the gala that joint work is essential for AI governance and other issues, and that while cooperation may not settle everything, nothing gets settled without it.
Based on reporting by CNBC Technology. RoxMag adds plain-English context and checks every figure against that source.
AI governanceUS-China relationsScott Bessentsuper intelligenceAI regulation
More in AI

Australia weighs criminal charges for knowingly selling unsafe AI
Canberra is considering a new offence for AI firms that release harmful models, following a review into an OpenAI data incident.

OpenAI builds Control Room kill switch for research models
OpenAI has deployed a central system that can halt output from research models within 35 seconds of a decision to stop them.

Lecturers and students disagree sharply on how much AI is used, study finds
Staff at one university estimated student AI use far higher than students did, according to research from Manchester and Basque academics.

Anthropic Moves Ahead With IPO Despite Market Turmoil
Anthropic will reportedly set a valuation range this week and start trading next week, as investors weigh the strength of the AI boom.