Yuyuan Tantian: Three Practical Directions for the U.S.-China AI Dialogue in November

Deep Tide TechFlow reports that as of October 2, the possibility of U.S.-China cooperation has become more optimistic compared to several months ago. Chinese negotiators have indicated that the two countries plan to address specific issues during the next dialogue in November. Specifically, there are at least three practical directions for advancement: First, jointly exploring AI risk classification and grading standards. Both sides need to clarify foundational issues first: What constitutes a major risk? Which model behaviors qualify as severe loss of control? These questions are prerequisites for subsequent discussions on more complex rules. Second, finalizing the scope, process, and requirements for reporting AI risk incidents. The U.S. and China have agreed to establish communication channels for AI-related incidents, marking the first step in addressing this gap. However, this mechanism cannot remain a verbal agreement. Clear constraints must be established regarding which incidents need to be reported, the timeframe for reporting, the information to be shared, and how to handle non-compliance with reporting requirements. Third, integrating AI governance into a broader framework of U.S.-China exchanges. Discussions on artificial intelligence should ensure full participation from China and amplify the voices of more countries. (Yuyuan Tantian) (Jin10)