Integration of Artificial Intelligence into Radio Access Network (AI-RAN): Architecture, Classification, and Development Focus

The incorporation of artificial intelligence (AI) into Radio Access Network (RAN) design is essential for network evolution toward the 6G future. This article provides a systematic categorization of AI-RAN into three primary approaches—AI-for-RAN, AI-and-RAN, and AI-on-RAN— according to functional level and application goals. Each methodology is elucidated structurally, supplemented by practical application examples and associated obstacles. The focal topics of several working groups aiding AI-RAN development are delineated, encompassing data administration, model lifespan, performance evaluation, security, and interoperability. The study findings indicate that AI can enhance spectrum efficiency, diminish latency, and facilitate intelligent services at the network's edge. The study concludes with prospective development pathways for constructing adaptive, autonomous, and secure RAN.