The evolution toward Sixth-Generation (6G) wireless systems introduces stringent requirements for ultra-low latency, high reliability, and massive scalability. Digital Twin Networks (DTNs) have emerged as a promising paradigm for real-time mirroring, simulation, and adaptive control of wireless infrastructure. However, existing studies often examine DTNs and Artificial Intelligence (AI) in isolation or limit their integration to narrow application scenarios, resulting in fragmented design and limited system intelligence. This survey addresses this gap by providing a comprehensive review of AI-enabled DTNs for wireless communications. It systematically examines the role of AI techniques, including deep learning, Reinforcement Learning (RL), Federated Learning (FL), and explainable AI (XAI), across the DTN lifecycle, encompassing twin creation, synchronization, prediction, decision-making, and feedback. A multidimensional taxonomy is proposed to classify AI-driven DTN approaches by functional objectives, network layers, and learning paradigms. In addition, the paper reviews state-of-the-art architectures, synchronization mechanisms, and modeling strategies, ranging from physics-based to hybrid AI-driven approaches. The analysis highlights how intelligent DTNs enable closed-loop control, scalable network orchestration, and enhanced anomaly detection, while also revealing open challenges related to generalization, interpretability, and energy efficiency. This survey provides a structured reference and research roadmap for advancing intelligent, trustworthy, and adaptive DTN-assisted 6G wireless networks.
Sanjalawe et al. (Tue,) studied this question.