Guest Editorial: Generative Foundation Models for Communications: Part 2

The evolution of wireless communication systems has been significantly influenced by conventional artificial intelligence (AI), which primarily contributes through the integration of algorithmic learning techniques with structured signal representations. By incorporating data-driven mechanisms that adapt protocols to diverse and dynamic wireless environments, traditional AI enhances performance in interference control, spectrum utilization, and mobility prediction. However, as wireless networks transition toward the sixth generation (6G), the increasing complexity brought by dynamic topologies, semantic-driven communication, large-scale edge intelligence, and environment-aware operations expose the limitations of traditional AI techniques. Specifically, they struggle to process high-dimensional and multi-modal environmental data, cope with decentralized and noisy conditions, and deliver real-time and context-sensitive intelligence across heterogeneous network infrastructures.