AI-Powered 6G: Technologies, Applications, and Challenges for Intelligent Connectivity

Abstract

The sixth-generation (6G) wireless paradigm represents a transformative leap in mobile communication, integrating Artificial Intelligence (AI) as a native capability across architecture, spectrum, security, and service layers. We expect 6G to meet the unmet demands for data rates, latency, connectivity, responsiveness, and intelligence. This paper presents a conceptual review of how Artificial Intelligence (AI) and Machine Learning (ML) serve as the foundational framework for embedding intelligence into 6G networks. This review examines how AI/ML is integrated into the necessary 6G requirements and its powerful technologies, such as Terahertz (THz) communication, Reconfigurable Intelligent Surfaces (RIS), and multi-domain integrated networks, and highlights why AI/ML is essential for optimisation and feasibility. It also discusses its key applications, such as massive MIMO optimisation, distributed edge AI, federated learning, autonomous UAV/V2X operations, intelligent network slicing, and the dynamic orchestration required for the massive Internet of Everything (IoE). Finally, it discusses significant open challenges and outlines key future research directions. The AI-native 6G boasts 10-fold reductions in latency, 50-fold increases in throughput, 2x energy savings, and near-99 % security accuracy over 5G. The future avenues are the transformative impact of AI-native 6G on digital twin-assisted control planes, THz-RIS co-design, and lifelong federated intrusion detection. These findings align with prior literature reporting gains in latency, throughput, security, and energy efficiency. They emerge together from a shared AI-driven orchestration layer spanning beamforming, federated security, and energy-aware computing. This distinguishes AI-native 6G from 5G-era approaches, where such optimisations were typically addressed independently rather than as part of a unified intelligent framework. To make AI-native 6G networks a reality, wireless networks should be transformed into fully intelligent and self-optimising systems. It must enable ultra-low latency, Tbps data rates, enhanced security, and energy-efficient operation across applications such as V2X, IoT, healthcare, and holographic communication. Overall, AI-native design makes 6G not just faster than 5G but autonomously adaptive and resilient, marking a paradigm shift toward networks that are inherently intelligent, sustainable, and capable of supporting future real-time digital ecosystems.