AI-Native 6G: The Future of Intelligent Connectivity

By Anselme Ndikumana Futureunmanned

 

AI-Native Networks Architecture (6G Vision) from white paper

 

A recent white paper from Khalifa University explores the concept of AI-Native 6G Networks, focusing on their architecture, embedded intelligence, and the path toward fully autonomous connectivity. The paper provides a compelling starting point for a clear, structured discussion of what it truly means for a network to be “AI-native.” Rather than treating artificial intelligence as an external optimization tool, it frames AI as a foundational element, deeply integrated into every layer of the network. This perspective not only redefines how future communication systems will be designed and operated, but also sets the stage for understanding the transformative potential of 6G in enabling intelligent, self-managing, and adaptive connectivity.

The evolution of mobile networks has never been about speed alone. From 1G to 5G, each generation has introduced transformative capabilities, improving connectivity, enabling new services, and reshaping how people and industries interact. However, 6G represents something fundamentally different. It is not simply “5G but faster.” Instead, it marks a paradigm shift toward AI-native networks by design.

What does AI-native mean for 6G?

In previous generations, artificial intelligence (AI) has been used primarily as an add-on to enhance performance, optimize resources, or automate specific tasks. In contrast, 6G networks are being envisioned with AI embedded at their core. This means intelligence is not layered on top; it is woven into every aspect of the system.

An AI-native network integrates sensing, reasoning, decision-making, and actuation across its entire lifecycle. The result is a system capable of autonomous, end-to-end operation, where networks can learn, adapt, and respond in real time without human intervention.

From automation to intelligence

The journey from 1G to 5G reflects a steady increase in software-driven capabilities and automation. With 6G, this evolution culminates in true network intelligence. The distinction between “AI-assisted” and “AI-native” becomes critical:

  • AI-assisted networks (5G and earlier): AI improves existing processes.
  • AI-native networks (6G): AI defines how the network operates from the ground up.

This shift introduces a new architectural concept: the AI plane, working alongside traditional control and user planes. AI functions are distributed across radio access networks (RAN), core networks, and edge environments, creating a deeply integrated intelligent ecosystem.

Enabling technologies behind AI-native 6G

Achieving this vision requires advances across multiple domains:

  • Distributed cloud-edge computing: Bringing intelligence closer to users for real-time decision-making
  • Data engineering: Efficient collection, processing, and sharing of massive datasets
  • Specialized AI hardware: Accelerating machine learning tasks within the network
  • Integrated sensing and communication: Allowing networks to perceive and understand their environment
  • Open architectures: Supporting flexible, plug-in AI modules and innovation

Together, these technologies form the foundation for a highly adaptive and intelligent network infrastructure.

Transformative capabilities

AI-native 6G networks will go far beyond current capabilities. Key functionalities include:

  • Self-optimization and self-healing: Networks automatically detect and fix issues
  • Predictive resource allocation: Anticipating demand before it occurs
  • Intent-based networking: Translating user needs into automated network actions
  • Hyper-personalized services: Tailoring experiences in real time

These capabilities represent a qualitative leap from 5G, enabling networks that are not just reactive but proactive and cognitive.

Real-world use cases

The potential applications of AI-native 6G are vast and transformative:

  • Ultra-immersive AR/VR experiences: Including real-time haptic feedback for the “metaverse”
  • Autonomous mobility systems: Vehicles and infrastructure collaborating seamlessly
  • Massive IoT and smart cities: Efficiently managing billions of connected devices
  • Industrial automation: Digital twins enabling real-time monitoring and optimization

These scenarios demand ultra-reliable, low-latency, and intelligent connectivity, precisely what 6G aims to deliver.

Rethinking performance metrics

Traditional metrics like throughput, latency, and reliability remain important, but they are no longer sufficient. AI-native networks introduce new key performance indicators, such as:

  • Learning accuracy of AI models
  • Decision latency in autonomous operations
  • Energy efficiency per AI inference
  • Level of network autonomy

These metrics help measure not just how fast a network is, but how intelligently it operates.

Security, trust, and governance

Embedding AI deeply into network infrastructure raises critical challenges:

  • Resilience to adversarial attacks targeting AI models
  • Data privacy and protection in large-scale data environments
  • Explainability and transparency of AI decisions
  • Regulatory oversight of autonomous systems

Addressing these concerns is essential to building trust in AI-native networks.

Business and operational impact

The transition to AI-native 6G will reshape the telecommunications industry:

  • Operators will shift toward automation-driven workflows
  • New value propositions and services will emerge
  • Collaboration between telecom companies, tech providers, and academia will intensify
  • Workforce skills will evolve, requiring expertise in AI, data science, and software engineering

This transformation will redefine how networks are built, operated, and monetized.

Looking ahead to 2030 and beyond

AI-native 6G represents a bold vision of the future, one where connectivity is autonomous, intelligent, and deeply integrated into every aspect of society. By embedding intelligence at its core, 6G will enable a new era of innovation, powering applications and services that are only beginning to be imagined.

As the roadmap unfolds toward 2030, the challenge lies not only in advancing technology but also in ensuring that these networks are secure, trustworthy, and aligned with human needs. If successful, AI-native 6G will not just connect the world, it will make it smarter.

Agentic AI Systems: What they are and what they Aren’t

By Anselme Ndikumana Futureunmanned

Evolution of AI (image from [1])

In the article “Agentic AI Systems: What It Is and Isn’t [1],” the authors  explore the emerging concept of Agentic AI and explain why it represents a new and powerful evolution of artificial intelligence.

They describe agentic AI systems as AI systems capable of autonomously pursuing complex goals over time. Unlike traditional AI tools that simply respond to prompts, agentic systems can plan, act, adapt, and make decisions in a more continuous and goal-driven manner.

At their core, agentic AI systems combine generative AI capabilities with a closed-loop architecture. This means they do not just generate outputs, they plan tasks, store and retrieve memory, use external tools, and continuously adapt based on feedback from their environment. They can decompose high-level objectives into sub-goals, learn from intermediate results, and adjust their strategies in real time.

To illustrate this, the authors provide examples such as a lifelong digital health coach that dynamically adjusts recommendations based on user behavior, or an intelligent travel concierge that not only suggests destinations but also plans, books, monitors changes, and automatically updates itineraries.

However, the authors also emphasize that increased autonomy brings new challenges. As AI systems begin to act more independently, important concerns arise regarding alignment (ensuring the system acts according to human intent), accountability (who is responsible for its decisions), and the risk of unexpected or emergent behaviors. Managing these risks is as critical as advancing the technology itself.

Another key issue highlighted in the article is conceptual confusion. The term “agentic AI” is often used loosely and sometimes conflated with chatbots, generative AI models, or traditional software agents. The authors argue that agentic AI constitutes a distinct category with specific technical and functional characteristics that must be clearly defined. Therefore, the article aims to bring conceptual clarity by:

  • Tracing the historical evolution of AI leading to agentic systems

  • Explaining how agentic AI differs from conventional AI agents and generative models

  • Exploring the opportunities and risks these systems present for businesses and consumers

Reference:                   

[1]. Dwivedi YK, Helal MY, Elgendy IA, Alahmad R, Walton P, Suh A, Singh V, Jeon I. Agentic AI Systems: What It Is and Isn't. Global Business and Organizational Excellence. 2026 Mar;45(3):253-63.

Towards AI-Native 6G: The Role of Large Language Models

By Anselme Ndikumana Futureunmanned

The evolution toward 6G networks marks a significant paradigm shift from static, rule-based architectures to adaptive, AI-driven network. At the forefront of this transformation are Large Language Models (LLMs), particularly Generalized Pretrained Transformers, which offer powerful capabilities for understanding user intent, generating action plans, and executing complex instructions. As such, LLMs are poised to become core enablers of next-generation networks and services. Recognizing this, the authors of this white paper [ 1] discussed AI-native 6G architecture, one that supports the seamless integration, provisioning, updating, and creation of diverse LLMs tailored to specific network functions and applications.

At the heart of this white paper [1]   is the concept of the AI-native 6G network that facilitates AI-centric operations across the network. This integrated approach promises transformative benefits such as Intelligent radio and network optimization, improving efficiency and adaptability, context-aware privacy and security.

Image from white paper:  Large language models in the 6G enabled computing continuum [1]

The convergence of 6G and advanced AI, embodied in scalable, responsive LLMs, will define the future of intelligent connectivity. The time to architect and invest in AI-native networks is now.

[1] Lovén, Lauri, Miguel Bordallo López, Roberto Morabito, Jaakko Sauvola, and Sasu Tarkoma "Large Language Models in the 6G-Enabled Computing Continuum: a White Paper (2025)"

 

Digital Twin Technology will improve 6G Networks, Study Reveals

By Anselme Ndikumana Futureunmanned

In a newly published article [1], researchers have outlined the critical requirements and capabilities of digital twin technology within 6G networks, highlighting sustainable deployment, real-time synchronization, seamless migration, predictive analytics, and closed-loop control.

The study also identifies emerging research opportunities that leverage both digital twins and artificial intelligence to optimize key aspects of 6G, including network performance, resource allocation, security, and intelligent service provisioning. By offering insights into how these technologies can work together, the authors discussed further innovation at the intersection of digital twin and 6G, ultimately paving the way for transformative applications and services in the near future.

[1] Liu W, Fu Y, Shi Z, Wang H. When digital twin meets 6G: Concepts, obstacles, and research prospects. IEEE Communications Magazine. 2024 Nov 4.

AI models for 6G wireless networks

By Anselme Ndikumana Futureunmanned
The authors in [1] discussed the Wireless Big Artificial Intelligence Models (wBAIMs).

Pre-training is a cornerstone feature of Big Artificial Intelligence Models (BAIMs), including their wireless variant, wBAIMs. This process eliminates the need for task- and scenario-specific training on targeted deployed devices. Instead, wBAIMs leverage pre-training, often through a collaborative effort between cloud and edge environments, to create versatile and efficient models ready for downstream applications.

The primary goal of pre-training in wBAIMs is to develop a generalized model that can be fine-tuned or prompted for specific wireless tasks and scenarios. This approach significantly reduces the complexity and computational overhead required for training on individual devices. By integrating pre-trained models, wBAIMs optimize their readiness for diverse applications, minimizing the time and resources needed for deployment.

A hallmark of the wBAIM-based architecture is its emphasis on integrating multiple wireless tasks into a unified framework rather than handling each task in isolation with separate models. Tasks such as:

  • Processing noisy reception pilots,
  • Managing compressed channel and signal sizes, and
  • Inferring user locations

are all seamlessly incorporated into the wBAIM. This integration showcases the model's ability to handle fundamental wireless functions cohesively.

The versatility afforded by wBAIM’s pre-trained architecture extends beyond basic tasks. By consolidating foundational wireless operations, wBAIMs pave the way for advanced applications across various domains. This holistic approach enhances system efficiency, enabling seamless support for complex and emerging wireless use cases.

The use of pre-training in wBAIMs not only optimizes their operational readiness but also aligns with the growing need for efficient, scalable solutions in wireless communications. As the technology evolves, wBAIMs are poised to revolutionize how wireless systems process, analyze, and adapt to dynamic scenarios, setting the stage for a new era in wireless intelligence.

This integration of pre-training strategies into the wireless domain underscores the potential of AI to innovate and streamline complex communication systems, ensuring robust performance across diverse applications.

 

[1]. Chen Z, Zhang Z, Yang Z. Big AI models for 6G wireless networks: Opportunities, challenges, and research directions. IEEE Wireless Communications. 2024 Jul 1.

logo

Free courses online

view all courses