By Viet Nguyen, President, 5G Americas (December 2024) –
It’s hard turning around nowadays without seeing a headline about AI. Indeed, AI is pervasive and its reach is gaining strength across an ever-expanding range of industries. Over the course of the past year, former 5G Americas President Chris Pearson wrote extensively about the topic in a five-part blog series focused on AI in wireless cellular networks. Today, I’m continuing in that trend, as 5G Americas is proud to release its most recent white paper Artificial Intelligence in Cellular Networks, which provides a much deeper examination of these transformative dynamics.
The ever-growing complexity of cellular networks has propelled AI into a pivotal role in telecommunications. As 5G evolves into 5G-Advanced and beyond, and as the industry looks toward 6G, AI’s integration across the network stack is revolutionizing operations, enabling automation, and enhancing user experiences.
Of course, nobody’s implementing AI just for the sake of it. There are real reasons to apply AI in our networks. Cellular networks are facing tremendous increasing demands, driven by the proliferation of connected devices and data-intensive applications. The latest Ericsson Mobility Report suggests mobile data growth increasing 21 percent annually and rising to 157 exabytes, as of Q3 2024. This growth renders traditional management approaches—relying on static rules and heuristics—insufficient. AI addresses these challenges with its capacity for real-time analysis, pattern recognition, and decision-making, ensuring seamless network performance and scalability. From optimizing resource allocation to enabling new network capabilities, AI is at the forefront of this technological evolution.
So how do we make wireless cellular networks AI-native?
Transitioning to AI-native wireless networks involves embedding AI as a core element across all layers and processes of the network. This includes leveraging cloud-native and distributed architectures, such as Open RAN and Cloud RAN, to enable scalable and flexible AI deployment. AI integrates into the physical, data link, and network layers to optimize operations like beamforming, mobility management, and traffic control while enabling cross-layer processes like intent-driven networking.
At the same time, robust MLOps ensures continuous training, monitoring, and updating of AI models, supported by high-quality, diverse datasets and federated learning for local adaptation. Trustworthy AI practices—transparency, bias mitigation, and privacy safeguards—are critical for maintaining reliability. Standardization by organizations like 3GPP and collaborative ecosystems further accelerates AI adoption. Ultimately, these advancements, combined with research into generative AI, advanced sensing, and next-gen computing, create a dynamic, adaptive infrastructure to support evolving network demands and innovations.
AI’s integration into cellular networks is layered, spanning from the physical layer (L1) to higher-level network management functions. At L1, AI optimizes the air interface, enhancing spectral efficiency and signal quality through advanced techniques like AI-enabled beam management and channel state feedback. For example, AI algorithms help compress and reconstruct high-dimensional channel state information, significantly reducing the overhead in 5G networks. These enhancements are crucial for improving the performance of technologies such as massive MIMO, a cornerstone of 5G deployments.
Here, you can see the kinds of AI-driven use cases in the physical layer:
At the data link (L2) and network layers (L3), AI contributes to mobility management, traffic optimization, and load balancing. By leveraging predictive analytics, AI can proactively manage handovers in dense urban environments, preventing disruptions in connectivity. Additionally, AI-driven strategies improve energy efficiency by dynamically deactivating underutilized cells during low-traffic periods—a capability that supports both operational cost savings and sustainability goals.
Beyond individual layers, AI also enables cohesive optimization through cross-layer processes such as intent-driven networking and lifecycle management. By understanding the network as a unified system, AI facilitates seamless collaboration across different layers and components, improving operational reliability and user experiences.
In particular, the Radio Access Network (RAN) is a focal point for AI-driven innovations. Open RAN architectures, which emphasize interoperability through standardized interfaces, benefit significantly from AI integration. RAN Intelligent Controllers (RICs) leverage AI for real-time optimization, supporting dynamic resource allocation and network slicing. These advancements enable more flexible and scalable networks, addressing the diverse requirements of 5G applications. In this figure below, you can see how AI workloads might interact across a variety of different RAN architectures.
Additionally, as AI becomes deeply embedded in cellular networks, ensuring its trustworthiness is paramount. Transparent and explainable AI systems build operator and user confidence. Key principles, such as robust monitoring, bias mitigation, and adherence to ethical guidelines, underpin responsible AI deployment. These measures ensure that AI-driven networks maintain integrity, safeguard user privacy, and operate reliably.
AI’s role will only expand as the industry transitions from 5G-Advanced to 6G. Future research in AI for telecommunications will focus on distributed architectures, cross-domain intelligence, and advancements in generative AI tailored to the industry’s unique challenges. These developments will underpin the next generation of services, from immersive applications like telepresence and digital twins to enhanced sensing and communication capabilities.
Collaboration among industry stakeholders—operators, technologists, and regulatory authorities—will be crucial in realizing the full potential of AI. Standardization efforts, such as those by 3GPP and the O-RAN Alliance, are already paving the way for interoperable and scalable AI solutions, and of course, the work never stops.
As you can see, artificial intelligence is reshaping cellular networks, driving efficiency, adaptability, and innovation. By integrating AI at every layer and across processes, telecommunications networks are evolving to meet the demands of a connected world. With its focus on responsible and strategic implementation, AI is not just a technological tool but a fundamental enabler of the telecommunications networks of the future.
As we move toward 6G, the importance of AI-driven innovation cannot be overstated. It promises to unlock unprecedented capabilities, ensuring that cellular networks remain at the heart of global digital transformation. Our trade association remains committed to helping you understand this incredible technology and how it will continue to resonate far into the future.
Keep reading. We will always have much more to say.
-Viet
