Generative AI-Assisted Management of Integrated Terrestrial and Non-Terrestrial Networks: A Comprehensive Survey

Generative Artificial Intelligence (GenAI) has recently emerged as a transformative paradigm capable of learning complex distributions, generating realistic data, and enhancing decision-making in uncertain environments. Originally developed for tasks in vision, language, and content creation, GenAI models, such as Generative Adversarial Networks (GAN), Variational Autoencoders (VAE), Generative Diffusion Models (GDM), and Large Language Models (LLMs), are now being extended to address challenges in communication networks. In particular, there is interest in using GenAI within the Integrated Terrestrial and Non-Terrestrial Network (ITNTN) paradigm, since the latter is envisioned as the backbone of 6G. This survey examines how GenAI can aid in managing ITNTN by analyzing recent research on GenAI-based solutions for resource allocation, task offloading, traffic prediction and generation, routing strategies, and channel modeling. We highlight the advantages brought by GenAI compared to traditional AI techniques, provide a structured taxonomy of existing work, and identify open challenges and future research directions. To the best of our knowledge, this is the first comprehensive study bridging the gap between GenAI and ITNTN, thus laying the ground for future intelligent and autonomous 6G network management.