AI to drive 6G evolution, but security and energy constraints threaten rollout

Despite the enormous potential described by the researchers, the study warns of major barriers that threaten the safe, sustainable and effective integration of AI in 5G and 6G systems. These obstacles span ethical, technical, operational and environmental dimensions, presenting the telecom industry with a complex balancing act.

Artificial intelligence (AI) is becoming the primary engine behind next-generation wireless systems, yet its integration into 5G and emerging 6G networks remains fragmented, uneven and slowed by structural, ethical, and interoperability challenges. The findings of a new paper published in Technologies suggest that while telecom operators, researchers, and policymakers are racing toward intelligent mobile networks, the transition to fully AI-enabled communications will require more coordination, governance, and infrastructure modernization than the industry has so far achieved.

The research, titled “Artificial Intelligence for 5G and 6G Networks: A Taxonomy-Based Survey of Applications, Trends, and Challenges”, offers a detailed taxonomy of technologies, operational impacts, use-case domains, integration models, and lifecycle considerations. The study’s core message is clear: the telecom ecosystem is undergoing a structural shift toward AI-native architectures, but obstacles ranging from energy constraints to privacy risks must be resolved before the full benefits of intelligent networks can be realized.

AI redefines the architecture and operations of 5G and 6G networks

The study highlights how 5G’s distributed, virtualized and cloud-native architecture has laid the groundwork for intelligent mobile systems. With increasing dependence on software-defined networking and network function virtualization, network operators have begun shifting away from hardware-centric design and toward scalable, AI-driven management. This transformation is accelerating as 6G research advances, bringing with it an expectation of pervasive machine intelligence embedded directly at the network core.

According to the researchers, AI is already playing a decisive role across multiple dimensions of mobile network operations. Machine learning and deep learning models support channel estimation, predictive optimization, mobility management, beamforming, congestion forecasting, resource distribution, anomaly detection, and real-time performance tuning. In fields such as IoT, smart cities, remote healthcare, autonomous transport, industrial monitoring, and extended reality applications, AI enables faster decision-making, higher service reliability and improved energy efficiency.

The study outlines a detailed classification system to show how AI functions across the entire network lifecycle. During the design phase, AI enables architecture planning, simulation, and configuration modelling. In deployment, AI supports automated placement, orchestration, and integration of virtual functions. During operation, it enables real-time monitoring, route optimization, traffic prediction, and interference mitigation. In maintenance, AI provides predictive analytics for fault detection, network healing, and lifespan extension. This lifecycle framing demonstrates how AI is no longer a supplementary tool but a core operational mechanism.

For 6G networks in particular, the authors emphasize that distributed intelligence will be essential. With future applications requiring near-instant processing, ultra-low latency, and adaptive reactivity, computation must move closer to the network edge. Edge AI, federated learning and real-time distributed models will become key enablers for autonomous vehicles, haptic communication, immersive virtual environments and large-scale machine-type communications. The survey stresses that these advanced uses cannot rely on cloud-only models; local intelligence embedded directly in devices, sensors and network nodes will be required for next-generation performance benchmarks.

AI’s expansion into terahertz communication, reconfigurable intelligent surfaces, network slicing, and massive MIMO systems marks a turning point in mobile technology evolution. As 6G ushers in unprecedented bandwidth and density, the network will need to operate as an adaptive, self-optimizing organism capable of learning and evolving over time.

Major challenges threaten progress toward intelligent, autonomous networks

Despite the enormous potential described by the researchers, the study warns of major barriers that threaten the safe, sustainable and effective integration of AI in 5G and 6G systems. These obstacles span ethical, technical, operational and environmental dimensions, presenting the telecom industry with a complex balancing act.

One of the most pressing concerns is data privacy. Intelligent networks depend on vast amounts of real-time data, including user locations, behavior patterns, device identifiers and environmental feedback. As the authors note, the distributed and continuous nature of this data collection raises risks around unauthorized access, inference attacks and loss of anonymity. Without robust privacy-preserving techniques such as federated learning, differential privacy and secure multi-party computation, future networks may expose users to unprecedented surveillance vulnerabilities.

The study also warns that interoperability remains a critical weakness across global mobile systems. Operators, vendors and researchers often work within isolated technological silos, creating fragmented approaches to AI integration. This lack of harmonization makes it difficult for AI models to communicate, scale and adapt across diverse infrastructures. The authors argue that international standards bodies must establish clearer guidelines to ensure consistent AI adoption and cross-vendor compatibility in the 5G–6G transition.

Another major issue is energy consumption. As AI workloads grow, the computational demands of training, inference and distributed learning increase sharply. This threatens the environmental sustainability of large-scale mobile networks. Edge devices, in particular, face power limitations that restrict their ability to run complex models. Without major advances in low-energy AI, neuromorphic computing, and hardware optimization, next-generation systems may become too resource-intensive to deploy at scale.

The researchers also highlight ethical concerns surrounding algorithmic transparency, fairness, and governance. In many cases, AI models used for network optimization are opaque, making it difficult for operators to identify errors, biases or unexpected behavior. The study argues that explainable AI is essential to building trust, ensuring accountability and preventing harm—particularly in safety-critical environments such as autonomous driving or remote healthcare.

Security vulnerabilities are another growing threat. AI-powered networks may face more sophisticated forms of adversarial attacks, data poisoning, model manipulation and distributed denial-of-service events. The authors emphasize that future telecom infrastructures must adopt multi-layered protection mechanisms that can detect anomalies, respond autonomously and adapt to evolving attack patterns.

Finally, the study warns about scalability constraints, noting that many AI solutions work well in controlled experiments but struggle under the complex demands of real-world networks. Ensuring that AI systems can scale across millions of devices and billions of data points will require significant advances in architecture design, orchestration logic and resource allocation.

The road ahead: A global shift toward ethical, scalable and distributed intelligence

The authors argue that the future of AI-powered mobile networks will depend on the industry’s ability to coordinate technological, regulatory and ethical frameworks at scale. A global shift toward distributed, secure and explainable intelligence will be essential to realizing the full potential of 6G.

Future networks will need to incorporate privacy-first intelligence, enabling data-driven decision-making without exposing sensitive information. Federated learning and encrypted computation will likely become standard tools for telecom operators. At the same time, explainable AI will emerge as a regulatory requirement, providing transparency for automated decisions that affect service quality, safety and user trust.

The researchers also foresee a future where multi-domain synergies become increasingly important. AI-enabled telecom systems will integrate with smart cities, healthcare platforms, autonomous transportation, industrial robotics, environmental monitoring and virtual reality ecosystems. This interconnected landscape will require new governance models that acknowledge the cross-sector nature of intelligent communication networks.

Scalability and energy efficiency will drive the next wave of innovation. Lightweight models, edge optimization techniques, and adaptive neural architectures will help address the growing carbon footprint of AI-powered connectivity. At the same time, advanced sensing technologies, intelligent surfaces and terahertz communication will expand the boundaries of what mobile networks can deliver.

A coordinated global strategy will be essential for building safe, sustainable and high-performance AI-enabled networks. Policymakers, researchers, and telecom operators will need to collaborate on standards, interoperability protocols, security requirements and ethical frameworks. The transition from 5G to 6G will be defined not just by faster speeds, but also by deeper intelligence, stronger safeguards and more human-centric governance