From zero-touch to intent-based networking: AI and LLMs for 5G and 6G network management

This tutorial provides an in-depth and practice-oriented overview of how Artificial Intelligence enables autonomous network management for 5G and future 6G systems. The tutorial starts by motivating the need for AI through the increasing complexity, heterogeneity, and performance requirements of next-generation networks, highlighting the limitations of manual and rule-based management. It introduces the concept of Zero-touch Service Management (ZSM) and presents a closed-control loop architecture integrating monitoring, analytics, and decision engines to achieve self-configuration, self-optimization, self-healing, and self-protection. Through concrete case studies, the tutorial demonstrates AI-driven solutions for different technological domains, including NFV lifecycle management (e.g., AI-based scaling of virtualized core functions), RAN self-configuration using deep reinforcement learning, and security management with ML-based DDoS detection and mitigation in network slicing contexts. The tutorial further explores emerging paradigms such as federated learning for privacy-preserving training and intent-based networking, where Large Language Models (LLMs) translate human-readable intents into actionable network configurations. It also addresses intent assurance using AI, XAI, and LLMs to guarantee SLA compliance. Overall, the tutorial bridges architectural concepts, algorithms, and real testbed implementations, showing why AI-native management is a cornerstone for scalable, flexible, and trustworthy 5G and 6G networks.