The AI Summit Between China and the United States: A Preview of the Interaction Between the Two AI Superpowers

According to monitoring by SenseTime Beating, the United States and China are preparing to launch a new official AI security dialogue, attempting to establish a "guardrail" for crisis prevention in their technological competition. The Wall Street Journal reported that the U.S. side will be led by Treasury Secretary Scott Bessent, while the Chinese side will involve officials such as Vice Minister of Finance Liao Min in preliminary mechanism discussions.

This official contact aims to establish a crisis management mechanism similar to that of the Cold War era. The core topics to be discussed include: preventing unforeseen behavior of AI models, controlling autonomous military systems, and addressing threats from non-state actors such as hackers who misuse powerful open-source tools. Industry speculation suggests that the two sides may explore the establishment of a high-level "AI hotline." Experts familiar with China's position point out that China maintains an open attitude towards discussing technology security and governance, with the core tone of its participation in the dialogue being "Stability, not alignment."

However, this dialogue aimed at prevention is facing a test of the U.S. lead negotiator's strong confrontational stance. According to reports from various external media outlets such as Bloomberg, Bessent has recently taken an extremely tough stance on the AI issue with China: in mid-April, he publicly stated that Anthropic's cutting-edge models would ensure U.S. dominance in the field of AI over China; just at the end of last month, he further criticized senators who invited scholars from Tsinghua University to participate in a U.S. AI security forum, asserting that "the real threat to AI security is allowing any country outside the U.S. to set global standards." This dual-track strategy, which seeks to prevent loss of control while rejecting other countries' involvement in rule-making, poses the biggest variable in these negotiations.

Historically, this is not the first time that the U.S. and China have attempted official AI interactions. The Biden administration facilitated the first round of high-level dialogue in 2023, but due to China's delegation of diplomatic officials rather than technical experts, substantive technical discussions were limited. However, outside of official contact, a "Track Two" (non-governmental) communication channel established by Henry Kissinger before his death in 2023 is still operational. Former Microsoft executive Craig Mundie and representatives from Tsinghua University and leading Chinese large-scale model companies continue to engage in private discussions on frontier model security and alignment issues.