https://ieee-icce.org/posts/openrit6g-icce-ai-native-6g-research-from-open-data-and-ai-pipelines-to-foundation-models-and-agentic-systems

OpenRIT6G@ICCE - AI-Native 6G Research: From Open Data and AI Pipelines to Foundation Models and Agentic Systems

Wednesday 29 July 2026

13:30 – 17:00

Organizers

Serge Fdida (Sorbonne Université)
Manu Gosain (Northeastern University)
Nguyen Huu Thanh (HUST)
Thomas Magedanz (TU Berlin)
Joyce Mwangama (University of Cape Town)
Nguyen Tai Hung (HUST)

Motivation

The emergence of AI-native 6G is reshaping the design, operation, and experimentation of future communication systems. Beyond AI-enabled optimization, future networks will increasingly rely on open research data, reproducible AI pipelines, foundation models, and agentic systems capable of supporting autonomous network operation.

The OpenRIT-6G initiative, formed in 2023 as an international workshop series for active researchers and developers, promotes an “open 6G for all”, which means that every country should be able to develop its own 6G infrastructure addressing the local specific needs and to develop local self-sovereign 6G eco ecosystems. OpenRIT-6G Research Infrastructures provide a unique open environment to investigate the above challenges through experimentally-driven research and practical demonstrations, enabling reproducibility, interoperability, and international collaboration.

This workshop will explore how open, modular 6G end-to-end architectures, open research data, MLOps frameworks, orchestrators, foundation models, and agentic AI can collectively support the development of AI-native 6G systems while addressing key concerns related to open source ecosystems, digital sovereignty, and international cooperation.

Programme

13:30 – 13:45

Welcome and Introduction

From 5G to AI-Native 6G: Why Open Experimental Research Matters

Nguyen Huu Thanh (HUST) & Serge Fdida (SU),

Topics:

AI-native 6G vision
Experimentation and reproducibility
Role of research infrastructures
International cooperation perspectives

13:45 – 14:15

Session 1

6G Challenges and Opportunities

Towards AI-Native 6G Networks

Thomas Magedanz TU Berlin

Topics:

6G vision, Open 6G
New services and opportunities
Experiences gained with the Fraunhofer FOKUS Open6GCore Toolkit

Q&A

14:15 – 14:45

Session 2

Understanding the Role of AI for network automation in emerging Low-Cost Open Source 5G/6G networks

Joyce Mwangama
University of Cape Town

· Understanding the complexity of emerging open 6G Networks

· Build your own Open Source low cost 6G end-to-end network

· Understanding the potential of AI for networks and AI in networks

· Introducing the new German-South African “AI4Open6GNet” Project

· The DIGITAfrica BluePrint

14 :45 – 15:15

Coffee Break

15:15 – 15:45

Session 3

MLOps and AI Experimentation Pipelines, From Data Acquisition to Continuous AI Deployment

Serge Fdida, Sorbonne Université[SF1]

Topics:

MLOps for networking
Data pipelines
AI lifecycle management
Experiment reproducibility
AI validation on SLICES-RI

15:45 – 16:15

Session 4

Agentic AI for 6G, Beyond Automation: Towards Autonomous Networks

Manu Gosain
Northeastern University

Topics:

Intent-driven networking
AI-assisted experimentation
AI-RAN: AI in RAN, AI on RAN, AI and RAN
What is AI-native networking?
Illustration with the Open 6G Initiative at NEU

Discussion:

Open-source versus proprietary AI ecosystems

16:15 – 17:00

Panel Discussion

Open Source, Sovereignty and International Cooperation for AI-Native 6G

Moderator
Serge Fdida (Sorbonne University)

Panelists

Joyce Mwangama (Cape Town University)
Thomas Magedanz (TU Berlin)
Manu Gosain (Northeastern University)
Nguyen Huu Thanh (HUST)
Nguyen Tai Hung (HUST)

Discussion Questions

Do we need telecom-specific foundation models?
What role should open source play in AI-native 6G?
Can AI-RAN remain reproducible without open research data?
How should Europe, Asia, Africa, and North America cooperate on AI-native 6G experimentation?
Can research infrastructures such as SLICES-RI become digital commons for AI-native networking?