Core Insights - The Shanghai Artificial Intelligence Security Work Committee was established to promote the synergy and governance innovation of "AI + security" technologies, addressing new threats such as deep forgery and data poisoning, and reshaping the industry's defense paradigm [1][2] - The committee aims to collaboratively develop standards and guidelines to mitigate security risks associated with large models across various dimensions [2] - The Shanghai cybersecurity industry has doubled since the beginning of the 14th Five-Year Plan, with a focus on the dual integration of "security + AI" and the development of new security industries in emerging fields [2] Group 1 - The Shanghai Artificial Intelligence Security Work Committee was established under the guidance of the Shanghai Municipal Economic and Information Commission, with representatives from multiple companies including Shanghai Yidian Group and SenseTime [1] - Hu Shaoyong, CTO of Shanghai Guanan, emphasized the need for a systematic and structural approach to prevent AI technology risks, highlighting the importance of security measures at each critical stage of a large model's lifecycle [1][2] - The committee's establishment is expected to enhance collaboration across the industry chain to address security risks effectively [2] Group 2 - The Shanghai Municipal Economic and Information Commission's Chief Engineer, Qiu Wei, noted that the cybersecurity industry in Shanghai has achieved significant growth and will focus on the dual integration of "security + AI" in future developments [2] - The Pudong New Area's "Mosu Space" and other platforms will be linked to create a collaborative development framework for the city's cybersecurity industry [2] - The Pudong District aims to integrate deeply into the national strategy for industrial innovation along the Shanghai-Nanjing corridor, establishing a regional cybersecurity "gravity field" [2] Group 3 - The 2024 Shanghai Cybersecurity Industry Innovation Achievements Directory was officially released, covering 20 innovative achievements across 14 directions, including AI security and quantum security [3]

世界互联网大会秘书长任贤良致辞。 南都讯记者程安带上VR眼镜，空间会议系统让教学资源变成三维动画；搭载自研芯片的最新款智能交 换机能够大幅增强安全性和可扩展性......5月23日，思科2025科技创新AI峰会在北京举行，大会以"智领 未来"为主题，汇聚了来自全球的行业专家、知名学者、企业领袖以及科技创新者分享最新的AI科技成 果和领先的行业经验，也展出了多样化的AI应用场景、高算力基础设施以及灵活的高性能AI网络解决 方案。 行业趋势：智算中心需求激增 AI从工具升级为核心引擎 去年11月，中国互联网网络信息中心发布的《生成式人工智能应用发展报告(2024)》显示，我国初步构 建了较为全面的人工智能产业体系，相关企业超过4500家，核心产业规模已接近6000亿元人民币，产业 链覆盖芯片、算法、数据、平台、应用等上下游关键环节。 但随着AI大语言模型的广泛应用，多模型叠加原来的混合多云架构带来了很多新的挑战和风险。思科 大中华区副总裁兼安全事业部总经理卜宪录向南都记者分析，目前企业对于AI应用在网络安全方向上 还处于赋能阶段，提升检测能力、使用更方便，但是在如何保护AI的基础架构和应用，比如识别网络 中的影子 ...