何卓丰,胡东滨,袁依格.大模型赋能网络安全:机遇与挑战[J].中国科学院院刊,2026,41(8):1671-1682.
大模型赋能网络安全:机遇与挑战
Large models empowering cybersecurity: Opportunities and challenges
大模型赋能网络安全:机遇与挑战
Large models empowering cybersecurity: Opportunities and challenges
作者
何卓丰1,2
中南大学 商学院 长沙 410083;湘江实验室 长沙 410205
HE Zhuofeng1,2
Business School, Central South University, Changsha 410083, China;Xiangjiang Laboratory, Changsha 410205, China
胡东滨1,2
中南大学 商学院 长沙 410083;湘江实验室 长沙 410205
HU Dongbin1,2
Business School, Central South University, Changsha 410083, China;Xiangjiang Laboratory, Changsha 410205, China
袁依格2*
湘江实验室 长沙 410205
YUAN Yige2*
Xiangjiang Laboratory, Changsha 410205, China
中南大学 商学院 长沙 410083;湘江实验室 长沙 410205
HE Zhuofeng1,2
Business School, Central South University, Changsha 410083, China;Xiangjiang Laboratory, Changsha 410205, China
胡东滨1,2
中南大学 商学院 长沙 410083;湘江实验室 长沙 410205
HU Dongbin1,2
Business School, Central South University, Changsha 410083, China;Xiangjiang Laboratory, Changsha 410205, China
袁依格2*
湘江实验室 长沙 410205
YUAN Yige2*
Xiangjiang Laboratory, Changsha 410205, China
中文关键词
网络安全;大模型;智能攻防体系;人工智能
英文关键词
cybersecurity;large models;intelligent cyber offense and defense systems;artificial intelligence
中文摘要
网络安全是国家安全与社会稳定的重要保障,网络安全大模型为安全攻防体系智能化提供了重要支撑。大模型技术作为人工智能领域先进的核心技术之一,为网络安全领域带来了新的应用研究方向。文章系统梳理了网络安全大模型的应用产品和应用场景,深入探讨网络安全大模型在模型能力、产业生态、可信安全等方面的发展趋势,以及在数据隐私、模型安全、更新维护、结果可信等方面面临的应用难点。研究认为,需要协调多方合作,推动情报共享、革新培养体系,锻造专业人才、构建模型生态,降低开发成本、完善政策标准,加强监督管理,以期加速网络安全技术与实践的智能化、现代化进程,助力构建网络空间的安全智能攻防体系。
英文摘要
Cybersecurity serves as a critical pillar for national security and social stability. Large models in cybersecurity are emerging as key enablers for the intelligent transformation of cyber offense and defense systems. As one of the most advanced core technologies in artificial intelligence, large models are introducing new research directions and application paradigms in the cybersecurity domain. This study systematically reviews the current landscape of cybersecurity-oriented large model applications and products, and explores their deployment scenarios in practice. It further analyzes the development trends in model capabilities, industry ecosystems, and trustworthiness, while identifying major practical challenges such as data privacy protection, model security, continuous updates, and result reliability. The study argues that coordinated multi-stakeholder collaboration is essential to advancing intelligence sharing, fostering new talent development frameworks, cultivating specialized expertise, building a model ecosystem, reducing development costs, improving policy and standards, and strengthening regulatory oversight. These efforts are crucial to accelerating the intelligent and modern evolution of cybersecurity technologies and practices, and to constructing secure and intelligent cyber offense and defense systems.
DOI10.3724/j.issn.1000-3045.20250228004

