Abstract: As a key infrastructure driving the development of the future digital society, the sixth-generation (6G) mobile communication system is undergoing deep integration with Large Language Model (LLM), propelling communication technology into a new stage centered on “native intelligence”. First, the revolutionary performance, architecture, and service scenario requirements of 6G compared to fifth-generation (5G) were systematically elaborated. The core roles and implementation pathways of LLM were focused on across four typical application scenarios: intelligent wireless resource management, network autonomy and intelligent maintenance, security and privacy enhancement, and support for emerging services. Then, the technical challenges faced by LLM in 6G applications were further examined, such as the balancing of model performance with privacy and security, computational resource constraints, and the complexity of cross-domain collaboration. Finally, future development trends were outlined, including evolution towards distributed, cloud-edge-end collaborative, cross-layer integrated security, and multi-technology synergistic frameworks, providing a systematic reference for research on the deep integration of 6G and LLM. 

Key words: Keywords: Sixth Generation (6G), mobile communication, large language model, native intelligence, wireless resource management, network autonomy, security and privacy, metaverse
 

摘要： 第六代（Sixth Generation，6G）移动通信系统作为引领未来数字化社会发展的关键基础设施，与人工智能大语言模型（Large Language Models，LLM）的深度融合正推动通信技术迈向以“内生智能”为核心的新阶段。首先，系统阐述6G在网络性能、架构及业务场景方面相较于第五代（Fifth Generation，5G）的革命性提升需求，重点剖析LLM在智能无线资源管理、网络自治与智能维护、安全与隐私增强以及新兴业务支撑等四大典型应用场景中的核心作用与实现路径。其次，进一步分析LLM在6G应用中面临的模型性能与隐私安全平衡、算力资源约束、跨域协同复杂性等技术挑战。最后，展望LLM向分布式、云边端协同、跨层融合安全及多技术协同演进的发展趋势，为6G与LLM的深度融合研究提供系统参考。

关键词: 6G, 移动通信, 大语言模型, 内生智能, 无线资源管理, 网络自治, 安全隐私, 元宇宙