Trustworthy, Explainable, and Sustainable Decentralized Intelligence for 6G Networks

Giovanni Perin; Michele Rossi; Enrique Tomás Martínez Beltrán; Fernando Torres-Vega; José María Jorquera Valero; Manuel Gil Pérez; Eunjeong Jeong; Nikolaos Pappas

cs.NISep 12, 2026

Trustworthy, Explainable, and Sustainable Decentralized Intelligence for 6G Networks

Authors: Giovanni Perin, Michele Rossi, Enrique Tomás Martínez Beltrán, Fernando Torres-Vega, José María Jorquera Valero, Manuel Gil Pérez, Eunjeong Jeong, Nikolaos Pappas,

Abstract

As 6G networks transition from theoretical frameworks into operational realities, artificial intelligence (AI) evolves from an add-on optimization tool into a distributed and interconnected structural layer. Unlike previous network generations that mostly relied on centralized cloud analytics platforms, AI-native 6G networks operate across a dynamic, multi-domain edge-cloud continuum where data originates from heterogeneous sources including user devices, radio access networks, sensing infrastructures, and vertical applications. Centralizing this massive volume of data creates severe communication overhead, unacceptable latency bottlenecks, single points of failure, and complex cross-domain governance challenges. Consequently, decentralization becomes a fundamental architectural requirement for future 6G network intelligence and zero-touch operations. Security serves as the primary enabler of this decentralized paradigm. Critical security functions, such as real-time threat detection, physical-layer attack mitigation, slice protection, and intrusion detection, require immediate access to local context and telemetry before operational data loses its value. However, moving intelligence to the edge via collaborative paradigms like federated learning (FL) and decentralized FL (DFL) introduces complex trade-offs. System security cannot be addressed in isolation; it is deeply intertwined with equally important aspects like trustworthiness, explainability, and energy sustainability. Taking these aspects into account, this paper develops a unified perspective on decentralized intelligence for 6G, arguing that decentralization, trustworthiness, explainability, and sustainability must be designed jointly rather than treated as independent requirements

Figures & tables

Fig. 1: Interdependent design dimensions of intelligence in AI-native 6G networks. Decentralization enables scalable distributed intelligence, privacy/security and explainability provide complementary reliability mechanisms, and sustainability ensures its operational feasibility.

Fig. 2: Evolution of AI paradigms from centralized to fully decentralized architectures.

Fig. 3: Schematic overview of the DFL closed-loop.

Role of XAI	Explanation signal	What it enables
Self-optimization	Feature and neuron attributions	Compact, drift-aware detectors tailored to edge constraints
Control-plane evidence	Machine-readable attribution record	Differentiated, automated response across xApp , rApp , and peers
Assurance	Attribution stability and integrity	Explanations that are robust, private, and faithful before policy acts on them

TABLE I: Three roles of explanations in a trustworthy, decentralized 6G IDS.

Fig. 4: On the left: input current

i(t)=1+A\sin\left(2\pi\frac{t}{B}\right)\exp\left(-C\,t\right)

(with

A=0.2

B=40

ms, and

C=0.01

ms -1 ), suppressed for

100\leq t\leq 125

and the corresponding membrane potential

u(t)

obtained from Eq. ( 3 ) and with the additional mechanisms of the firing and reset thresholds. Here, we used

\tau_{m}=20

ms. On the right: schematic representation of a spiking neuron with three inputs (multiplied by the network weights) and the resulting output current.

Giovanni Perin ([email protected]) received his Ph.D. degree in Information Engineering from the University of Padova, Italy. He is an Assistant Professor at the University of Brescia, Italy. His research focuses on distributed learning, optimization, and sustainable AI and networks. He is a Member of IEEE.

Table 6

Michele Rossi ([email protected]) received his Ph.D. degree in Information Engineering from the University of Ferrara, Italy. He is a Full Professor at the University of Padova, Italy. His research focuses on sustainable AI , wireless sensing, and next-generation networking. He is a Senior Member of IEEE.

Table 7

Enrique Tomás Martínez Beltrán ([email protected]) received his Ph.D. degree in Computer Science from the University of Murcia, Spain, where he is currently a Postdoctoral Researcher within the CyberDataLab. His research focuses on collaborative learning, large language models, and cybersecurity.

Table 8

Fernando Torres-Vega ([email protected]) is a B.Sc. student in Cybersecurity at the International University of La Rioja, Spain. He is currently a software developer at the University of Murcia, Spain. His research interests include distributed systems, decentralized architectures, and cybersecurity.

Table 9

José María Jorquera Valero ([email protected]) received the M.Sc. and Ph.D. degrees in Computer Science from the University of Murcia. He is currently a Postdoctoral Researcher with the CyberDataLab, University of Murcia. His scientific research interests include trust management, cybersecurity, 5G networks, intent-based management, and continuous authentication.

Table 10

Manuel Gil Pérez ([email protected]) received the M.Sc. and Ph.D. degrees in Computer Science from the University of Murcia. He is an Associate Professor with the Department of Information and Communication Engineering, University of Murcia, Spain. His scientific activity focuses mainly on cybersecurity, trust and reputation management, and security operations.

Table 11

Eunjeong Jeong ([email protected]) received her Ph.D. degree in Communication Systems from EURECOM, France. She is currently a Postdoctoral Researcher at Linköping University, Sweden. Her research focuses on federated learning and energy-efficient scheduling. She is a Member of IEEE.

Table 12

Nikolaos Pappas ([email protected]) received his Ph.D. degree in Computer Science from the University of Crete, Greece. He is an Associate Professor at Linköping University, Sweden. His research interests include semantic communications, age of information, network-level cooperative networks, and performance analysis and stochastic modeling. He is a Senior Member of IEEE.

Table 13

Farah Abed Zadeh ([email protected]) received her M.Sc. degree in Data and Computational Science from University College Dublin, Ireland, where she is currently a Ph.D. student in Computer Science. Her research focuses on explainable AI in next-generation networks.

Table 14

Chamara Sandeepa ([email protected]) is a Postdoctoral Research Fellow at University College Dublin, Ireland. His research interests include AI security, explainable AI, federated learning, and intelligent security solutions for next-generation networks.

Table 15

Bartlomiej Siniarski ([email protected]) received his Ph.D. degree from University College Dublin, Ireland. He is a Researcher and Co-Director of the Network Softwarization and Security Labs (NetsLab), and is involved in several EU-funded research projects. His research interests include 5G/6G networks, IoT, network security, and network management.

Table 16

Madhusanka Liyanage ([email protected]) is a Professor and Ad Astra Fellow at the School of Computer Science, University College Dublin, Ireland, and Director of the Network Softwarization and Security Labs (NetsLab). His research interests include 5G/6G security, artificial intelligence, explainable AI, federated learning, blockchain, and edge computing. He is a Senior Member of IEEE.

Table 17

Betül Güvenç Paltun ([email protected]) is a Senior Researcher at Ericsson Research, Turkey. Her research focuses on trustworthy AI and intrusion detection systems.

Table 18

Leyli Karaçay ([email protected]) received her Ph.D. degree in Computer Science and Engineering from the University of Sabanci, Turkey. She is a Senior Security Researcher at Ericsson Research, Turkey. Her research focuses on distributed learning, network security, and trustworthy AI.

Table 19

Ioannis Pitsiorlas ([email protected]) received his M.Sc. degree in Data Science and Engineering from EURECOM, France, where he is currently a Ph.D. candidate. His research focuses on trustworthy and explainable AI.

Table 20

Marios Kountouris ([email protected]) received the Ph.D. degree in Electrical Engineering from Télécom Paris, France. He is a Full Professor at EURECOM, France, and a Distinguished Researcher at the University of Granada, Spain. His research focuses on communications theory and machine learning for communications. He is a Fellow of IEEE.

Table 21

Trustworthy AI

See all 28 papers

Towards Resilient and Autonomous Networks: A BlueSky Vision on AI-Native 6G
May 20, 2026Liang Wu, Kelly Wan, Mayank Darbari +16G Networks Multi-Agent Systems
Nokia Sunnyvale, California, USA
6G Needs Agents: Toward Agentic AI-Native Networks for Autonomous Intelligence
May 2, 2026Mohamed Amine Ferrag, Abderrahmane Lakas, Merouane Debbah6G Networks LLM Agent Orchestration
Department of Computer and Network Engineering, United Arab Emirates University, UAE · Research Institute for Digital Future, Khalifa University, UAE
Heterogeneity-Aware Belief Synchronization for Semantic Communication in AI-Native 6G Networks
Aug 13, 2026Muhammad Hannan Akram, Muhammad Abubakar Rashid, Wassi Haider Kabir +36G Networks Semantic Communication
School of Electrical Engineering and Computer Science (SEECS), National University of Sciences and Technology (NUST), Islamabad 44000, Pakistan · Department of Electronic Engineering, Kyung Hee University, Yongin 17104, South Korea · CONNECT Centre and the Department of Computer Science, Munster Technological University, Ireland

May 20, 2026J/K move · Enter open · S save

Towards Resilient and Autonomous Networks: A BlueSky Vision on AI-Native 6G

Liang Wu, Kelly Wan, Mayank Darbari, Liangjie Hong

Nokia Sunnyvale, California, USA

6G Networks Multi-Agent Systems

Open paper AskPDF arXiv

The proliferation of emerging applications, such as autonomous driving and immersive experiences, demands cellular networks that are not only faster, but fundamentally more resilient and autonomous. This paper presents a BlueSky vision on how Artificial Intelligence will be natively integrated into 6G, shifting the paradigm from Network for AI to AI for Network. We envision that, unlike 5G's reliance on scattered, ad-hoc models each trained for a single task, native AI in the 6G era will be anchored by a foundation model and and orchestrated via collaborative multi-agent systems, framing network management as a unified, multi-modal, multi-task optimization problem. Built on this vision, we outline two transformative directions. The first focuses on developing a 6G foundation model as a unified backbone, with task-specific knowledge distilled into compact models suited for diverse edge deployments. The second advances multi-agent systems designed to autonomously diagnose, maintain, and recover networks with minimal human intervention. These directions chart a roadmap for 6G to evolve into an intelligent, self-sustaining communication infrastructure.

Abstract

Trustworthy AI

Explore similar work

Towards Resilient and Autonomous Networks: A BlueSky Vision on AI-Native 6G

6G Needs Agents: Toward Agentic AI-Native Networks for Autonomous Intelligence

Heterogeneity-Aware Belief Synchronization for Semantic Communication in AI-Native 6G Networks