The United States and China have agreed to establish a communication channel for dealing with artificial intelligence incidents, as the world's two leading AI powers seek to improve coordination over emerging technology risks.
The agreement was announced by the White House after US President Donald Trump and Chinese President Xi Jinping held talks focused partly on the growing challenges associated with artificial intelligence.
According to a White House fact sheet, the two countries will also establish a US-China “Super Intelligence” dialogue to discuss the potential risks and benefits of advanced AI systems. The White House said the two sides agreed to use the term “Super Intelligence” rather than AI in the dialogue.
The next exchange under the new dialogue is expected to take place by November. The separate bilateral communication channel is intended to provide a way for the two governments to communicate when AI-related incidents occur.
The initiative follows earlier discussions between US Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng. During those talks, the two sides discussed an “incident line” for communicating about AI safety emergencies and possible protocols for handling major risks.
Bessent previously said the discussions could cover risks such as uncontrollable AI agents, cyberattacks carried out by non-state actors and other incidents that could have national security implications. He also argued that greater transparency between the two major AI powers was important.
The talks come as both countries compete for leadership in advanced AI while facing shared concerns about cybersecurity, loss of control and the possible misuse of increasingly capable systems. Recent AI-related cyber incidents have added urgency to international discussions about safety and accountability.
Despite the new communication arrangement, significant differences remain between Washington and Beijing over technology, national security and access to advanced computing and AI technologies. The proposed channel therefore represents a limited area of cooperation within a broader strategic rivalry.
The agreement marks a step toward direct communication on AI emergencies, but its effectiveness will depend on how the mechanism is implemented and whether officials on both sides can maintain communication during a serious incident.