Earlier today, the Frontier AI and Agent Security Forum was held in Shanghai during the 2026 World Artificial Intelligence Conference. Over 20 global experts and policymakers jointly released multiple reports and achievements in the field of AI safety, and engaged in discussions on frontier AI risk management and global industrial governance. The forum focused on three core themes: "Frontier AI Risk Monitoring," "Agent Out-of-Control Risks and Emergency Management," and "Frontier AI Risk Management with a Global Perspective." It featured keynote speeches and roundtable discussions involving representatives from United Nations agencies, leading AI enterprises worldwide, international think tanks, and domestic research institutions. A total of five landmark research outcomes were unveiled at the event, covering technical tools, tiered standards, emergency response frameworks, and industry-wide scenario assessments, collectively forming a comprehensive AI security management system.