Introduction

Treasury Secretary Scott Bessent delivered a direct message to the National Committee on U.S.-China Relations: Washington and Beijing can work together to manage the dangers of artificial intelligence, but neither side should pretend their rivalry has disappeared. Speaking in New York, Bessent framed the relationship as a blend of competition and cooperation, warning that the two powers must balance their competitive instincts with the responsibilities of leading a transformative technology.

What Happened

Bessent announced that the United States and China have created a new dialogue channel focused on what he called "super intelligence," or SI. The mechanism is intended to reduce shared risks and prevent misunderstandings between the two countries' political systems. Under the arrangement, both governments would exchange best practices for testing advanced AI models before public release, stress that human judgment must remain in control of the technology, and notify each other about serious incidents involving powerful SI systems.

Despite the new channel, Bessent said no commitment has been made to slow or cap AI development. The United States and China remain locked in what he described as an intense race for global leadership in artificial intelligence. The Trump administration has generally resisted wider restrictions on AI research, partly out of concern that limiting U.S. developers would hand an advantage to China. President Trump has indicated some openness to regulatory safeguards, but only if they do not leave the United States in what he called a "deep second position."

The remarks came ahead of an expected meeting between the two countries' leaders at the Asia-Pacific Economic Cooperation summit in November. Bessent is expected to join that delegation. Washington and Beijing had previously agreed during Chinese President Xi Jinping's September visit to establish an AI communications channel intended to defuse serious incidents.

Why This Matters

The announcement highlights a delicate balancing act. Both governments recognize that unchecked AI development could create dangers that cross borders, yet neither wants to fall behind in a strategic technology race. By creating a risk-focused dialogue without agreeing to limits on advancement, the U.S. and China are attempting to manage the hazards of AI while preserving their freedom to compete.

The timing adds urgency. The U.S. has already restricted exports of its most advanced semiconductors to Beijing and has accused Chinese developers of replicating American models through large-scale distillation. Earlier this year, access to certain Anthropic models was briefly suspended for foreign nationals before being restored. At the same time, the upcoming APEC summit is clouded by corporate controversy, including reports surrounding Nvidia CEO Jensen Huang that have raised questions about his participation in the presidential delegation.

Key Takeaways

  • The U.S. and China have launched a new AI dialogue focused on "super intelligence" risks and incident prevention.
  • Both sides will share testing best practices and agree to keep human judgment central to advanced AI systems.
  • No agreement exists to limit or slow the development of artificial intelligence.
  • The Trump administration remains cautious about AI regulation that could weaken U.S. competitiveness against China.
  • Leaders from both nations are expected to meet again at the November APEC summit, with Bessent attending.

Conclusion

Bessent's speech makes clear that cooperation on AI safety does not mean the end of U.S.-China rivalry. The two countries are building a narrow channel to exchange warnings and testing methods, but they are not promising to stop racing toward more powerful systems. As the November summit approaches, the real test will be whether this dialogue can prevent dangerous misunderstandings while both governments continue to pursue AI dominance.