On September 15, the 5th Beidou Large-Scale Application International Summit commenced in Zhuzhou, Hunan Province. Zhou Hongyi, the founder of 360 Group, delivered a keynote address as the sole invited representative from a cybersecurity firm. He emphasized that as the scope of Beidou applications broadens, the significance of a secure foundation becomes ever more paramount. He advocated for the adoption of the 'governing models with models' approach to establish a formidable security defense system.

Zhou Hongyi noted that Beidou has been seamlessly integrated into a multitude of industries, transitioning from isolated applications to widespread adoption, and evolving from a mere navigation tool into a comprehensive spatiotemporal infrastructure. With the rapid advancement of AI, the fusion of 'large models + intelligent agents' has emerged as a novel paradigm for Beidou applications. However, this fusion also presents threefold challenges: AI-driven automated vulnerability detection leading to a proliferation of attack vectors, expedited attack chains, and the inherent risk of AI malfunctioning.

To address these challenges, he put forward three recommendations: leveraging intelligent agents to automatically detect and rectify vulnerabilities, employing intelligent agents for defense to achieve 'self-driving' cybersecurity, and utilizing third-party security intelligent agents to oversee the operations of work intelligent agents, ensuring their behavior remains controllable. 360 has already rolled out two major AI security products, 'Yitian' and 'Tulong', which have been deployed in over 500 enterprises and have successfully identified more than 10,000 vulnerabilities.