This round of dialogue focused on clarifying terminology related to "loss of control," AI risk stratification, and security threats posed by the misuse of AI by non-state actors. Participants also exchanged views on the profound impact of AI on nuclear and biological security, as well as AI alignment strategies to prevent loss of control, aiming to further enhance mutual awareness and understanding. Additionally, experts from both sides shared updates on AI policy trends in China and the United States and explored practical pathways to mitigate the risks of AI systems losing control.

Moving forward, both project teams plan to maintain candid and in-depth exchanges in line with the latest trends in AI technology. The two sides will continue to jointly address AI security risks and challenges, striving to reach further consensus and achieve more substantive outcomes for the project.