Toward Seamless 6G and AI/ML Convergence: Architectural Enhancements and Security Challenges

Wei Yi, Yulong Fu, Jin Cao, Lu Gan, Lihui Xiong, Hui Li

Abstract: As mobile communication networks evolve toward 6G, integrating Artificial Intelligence (AI) and Machine Learning (ML) becomes crucial to enhance network efficiency, automate operations, and enable advanced applications such as autonomous systems and smart cities. However, current integrations predominantly focus on the business layer and fail to fully exploit the vast amounts of high-quality terminal data and available computational resources. To address these limitations, we delineate the systemic integration of 6G and AI into three pivotal stages and evaluate how current standards align with these stages. Next, we propose an evolutionary 6G AI/ML security architecture built on the existing 5G system. It can achieve seamless integration by enhancing network functions and reusing communication interfaces. We further illustrate this integration through two innovative use cases at the system level, demonstrating the architecture’s robust capabilities in data management, privacy protection, and communication interfaces. We also identify potential security challenges inherent in integrating 6G and AI/ML, and provide targeted suggestions. Following 3GPP standards, this work aligns with the 6G evolutionary trajectory and can lay a foundational framework for future 6G development.

External IDs:dblp:journals/network/YiFCGXL25