The surge in connected devices in 6G with typical massive access scenarios, such as smart agriculture, and smart cities, poses significant challenges to unsustainable traditional communication with limited radio resources and already high system complexity. Fortunately, the booming artificial intelligence technology and the growing computational power of devices offer a promising 6G enabler: semantic communication (SemCom). However, existing deep learning-based SemCom paradigms struggle to extend to multi-user scenarios due to their rigid end-to-end training approach. Consequently, to truly empower 6G networks with this critical technology, this article rethinks generative SemCom for multi-user system with multi-modal large language model (MLLM), and propose a novel framework called "M2GSC". In this framework, the MLLM, which serves as shared knowledge base (SKB), plays three critical roles for complex tasks, spawning a series of benefits such as semantic encoding standardization and semantic decoding personalization. Meanwhile, to enhance the performance of M2GSC framework and to advance its implementation in 6G, we highlight three research directions on M2GSC framework, namely, upgrading SKB to closed loop agent, adaptive semantic encoding offloading, and streamlined semantic decoding offloading. Finally, a case study is conducted to demonstrate the preliminary validation on the effectiveness of the M2GSC framework in terms of streamlined decoding offloading.

翻译：随着6G网络中连接设备数量的激增，典型的大规模接入场景（如智慧农业和智慧城市）对传统通信模式提出了严峻挑战，后者受限于无线电资源且系统复杂度已居高不下，难以持续发展。幸运的是，人工智能技术的蓬勃发展和设备计算能力的不断提升，为6G提供了一个极具前景的使能技术：语义通信。然而，现有基于深度学习的语义通信范式因其僵化的端到端训练方式，难以扩展到多用户场景。因此，为了真正将这一关键技术赋能于6G网络，本文重新思考了面向多用户系统的多模态大语言模型生成式语义通信，并提出了一种名为"M2GSC"的新颖框架。在该框架中，作为共享知识库的多模态大语言模型，在复杂任务中扮演着三个关键角色，从而催生出一系列优势，如语义编码标准化和语义解码个性化。同时，为了提升M2GSC框架的性能并推动其在6G中的实施，我们重点指出了围绕该框架的三个研究方向：将共享知识库升级为闭环智能体、自适应语义编码卸载以及流线化语义解码卸载。最后，通过一个案例研究，初步验证了M2GSC框架在流线化解码卸载方面的有效性。