The Role of Large Language Models in Designing Reliable Networks for Internet of Things: A Short Review of Most Recent Developments
Melchizedek Alipio, Miroslav Bureš · IEEE Access · 2025
The rapid growth of the Internet of Things (IoT) networks has increased the need for intelligent, flexible, and scalable networking solutions. This paper reviews the use of Large Language Models (LLMs) in improving network protocols, automating decision-making, and strengthening security in IoT networks. A detailed analysis is conducted to classify existing research based on applications, network types, methodologies, and performance metrics. LLMs have been used in network configuration, security monitoring, cyber threat detection, federated learning, and improving network performance. Their integration with edge computing, 6G networks, and AI-driven network control enables real-time network adjustments, automated troubleshooting, and efficient traffic management. However, challenges like high computing demands, energy consumption, security risks, and slow adaptation in dynamic networks still exist. This paper identifies emerging trends, including LLM-based self-learning networks, privacy-aware AI training, and hybrid AI models that combine graph-based neural networks, reinforcement learning, and multimodal AI. By reviewing recent research from 2023 to early-2025, this study provides a clear understanding of how LLMs transform IoT and network management. The discussion highlights future research directions, focusing on decentralized AI frameworks, optimized model training, and AI-driven network automation, aiming to develop more efficient, secure, and reliable network infrastructures.