U.S. Treasury Secretary Scott Bessent has cautioned that Washington and Beijing should “harbor no illusions” about their fundamental differences as they engage in managing artificial intelligence risks while continuing to compete for technological supremacy. Speaking to the National Committee on U.S.-China Relations in New York on Tuesday, Bessent outlined the framework for a newly established dialogue mechanism focused on “super intelligence,” or SI, designed to mitigate shared risks and prevent misunderstandings between the two systems.
The proposed framework envisions an exchange of “best practices” for testing advanced AI models prior to their public release, with a focus on maintaining human oversight of the technology. Both nations are expected to alert one another regarding serious incidents involving these models. However, Bessent did not commit to imposing specific limits on the advancement of AI technology, reflecting the ongoing intense race for global dominance in the sector.
The Trump administration has generally resisted broad regulatory restrictions on AI development to avoid ceding a competitive advantage to China. While President Donald Trump recently expressed some openness to regulatory safeguards, stating he “doesn’t mind guardrails,” he emphasized that regulations should not place the United States in a “deep second position” relative to China. Conversely, Chinese officials, including the foreign ministry and think tank founder Huiyao Wang, have dismissed calls to slow down frontier models, arguing that such fearmongering disrupts global governance and that risks should be managed cooperatively.
Washington and Beijing agreed to launch this AI dialogue during President Xi Jinping’s visit in September, establishing a communications channel to help defuse serious incidents. The two leaders are scheduled to meet for the third time next week on the sidelines of the Asia-Pacific Economic Cooperation Leaders’ Meeting on November 18 and 19, with Bessent attending as part of the U.S. delegation.
Good that they talk, but no real limits means we are still racing toward the edge.
Honestly, I don’t trust them to agree on anything this important without it causing more problems.