AI-Native Radio Access Networks in the UK: From Static Radio Resource Management to Predictive 5G Optimisation

The continued growth of mobile data traffic is placing increasing demands on radio access networks (RANs).Although 4G and 5G networks already use sophisticated automation to manage radio resources, conventionaloptimisation methods can struggle to respond efficiently to rapidly changing traffic patterns, complex interferenceconditions and variations in user behaviour. Artificial intelligence (AI) and machine learning (ML) offer an opportunityto make network management more predictive by identifying patterns in network data, forecasting demand andsupporting more adaptive resource allocation. This article examines the development of AI-native RAN, focusing onthe UK 5G environment. It reviews radio resource management, predictive optimisation and emerging AI-RANarchitectures, including the O-RAN Intelligent Controller framework. It considers the O2–Nokia live deploymentannounced at London Paddington in October 2026 and BT's planned AI-RAN trial with Nokia and NVIDIA inBirmingham during 2027. It also examines data quality, model reliability, security, interoperability, computingoverhead and human oversight. The article argues that AI-native RAN should be treated as an evolution ofestablished radio engineering, with benefits demonstrated through measurable improvements in network efficiency,reliability and customer experience.