SINGAPORE — The United States and China have agreed to continue discussions on artificial intelligence safety despite intensifying competition between the two countries in advanced technology. US Treasury Secretary Scott Bessent said the bilateral AI dialogue had been established after his meeting with Chinese economic official He Lifeng in New York. He said the next meeting could take place in about two months.
Washington has also proposed a notification system for AI-related incidents that could threaten national security. The plan could include designated points of contact, procedures for classifying incidents and emergency communication channels, although details have not been disclosed.
Analysts said the agreement was a limited but significant step because it showed that both sides recognised the risks posed by AI, including cybersecurity incidents, AI weaponisation and the possibility of losing human control over advanced systems.
The United States and China are developing some of the world’s most powerful AI models, but their rivalry has been strained by accusations that Chinese companies copied or extracted capabilities from leading American systems. Beijing has rejected the allegations as groundless.
Experts said the proposed notification mechanism could help prevent miscalculations during an AI-related crisis, but questioned whether it would be used effectively. They also warned that competition between private companies and broader disputes over technology restrictions would continue to limit cooperation.
The planned dialogue follows earlier discussions between US President Donald Trump and Chinese President Xi Jinping about creating a formal channel for AI cooperation. Analysts said future progress would likely depend on building trust through limited areas of shared interest rather than attempting to resolve the wider technology rivalry.