US and China discuss AI safety plan ahead of Trump-Xi summit

Top officials from the United States and China have engaged in discussions regarding artificial intelligence (AI) safety, including the proposal of a new "notification mechanism" for incidents that could pose national security threats. These preparatory talks, held in New York between US Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng, are setting the stage for an upcoming summit between US President Donald Trump and Chinese President Xi Jinping in Washington this week. The proposed system aims to enhance transparency and communication between the two leading AI powers, particularly concerning national security-level AI events.

Both nations have agreed to establish a formal "US-China AI dialogue" and intend to reconvene in approximately two months to continue these discussions. While US officials characterized the initial engagement as successful, Chinese state media offered a more concise acknowledgment of AI discussions, indicating potential differences in emphasis. Analysts suggest that the forthcoming summit is likely to yield modest advancements rather than significant breakthroughs, but the commitment to ongoing dialogue on AI safety is considered a notable development.

These AI safety conversations are part of broader discussions between the two countries, which also encompass issues such as trade tariffs and critical minerals. The initiative to create an AI communication channel highlights a mutual recognition of the urgent risks associated with artificial intelligence, even amidst intense technological competition and existing mistrust.