On July 19, amid the grand occasion of the 2026 World Artificial Intelligence Conference, Shanghai played host to the Frontier AI and Agent Safety Forum. This pivotal event saw the convergence of over 20 preeminent global experts and policymakers, who assembled to unveil a series of groundbreaking reports and advancements in the realm of AI safety. The forum served as a fertile ground for profound discussions on cutting-edge AI risk management and the global governance landscape of the industry.

The forum's agenda was structured around three pivotal themes: 'Monitoring Risks in Frontier AI,' 'Mitigating Agent Runaway Risks and Emergency Response,' and 'Global Perspectives on Frontier AI Risk Management.' Through a series of insightful keynote speeches and engaging roundtable dialogues, the forum successfully bridged the gap between senior practitioners from pertinent UN bodies, leading AI enterprises both domestically and internationally, esteemed international think tanks, and top-tier experts from domestic research institutions. This amalgamation provided a unique platform for the interplay of technical expertise and policy deliberation.

Furthermore, the forum marked the debut of five landmark research achievements. These encompassed a range of crucial elements, including technical tools, classification criteria, emergency response strategies, and a holistic industry analysis. Together, these elements forged a comprehensive and robust AI safety management framework, setting a new benchmark for the industry.