Khalifa University of Science and Technology has launched a groundbreaking artificial intelligence model, RF-GPT, designed to interpret wireless signals—marking a major leap in how telecom networks could be managed and optimised.
Developed by the university’s Digital Future Institute, the new system is being described as the first AI “language model” capable of understanding radio-frequency (RF) signals, a domain traditionally beyond the reach of conventional AI tools.
A breakthrough beyond text-based AI
Unlike typical AI models that process text or structured data, RF-GPT can analyse wireless signals by converting them into visual patterns known as spectrograms. These patterns are then interpreted by the model, allowing it to “understand” activity within the electromagnetic spectrum and respond in plain language.
In testing, RF-GPT delivered standout results—outperforming existing baseline models by up to 75.4% in RF-related tasks and accurately counting signal sources nearly 98% of the time.
Powering next-generation networks
Researchers say the innovation could transform how telecom operators and regulators monitor and manage increasingly complex wireless environments, from Wi-Fi networks to 5G infrastructure.
The model can:
- Identify different types of signals
- Detect overlapping transmissions
- Recognise wireless standards
- Estimate device activity in networks
- Extract data from advanced systems like 5G
The project aligns with the UAE’s broader push toward AI-driven infrastructure under its national artificial intelligence strategy, laying the foundation for more autonomous and intelligent wireless systems.
Built by global research collaboration
The project was led by Merouane Debbah, Senior Director at the Digital Future Institute, alongside an international team of researchers from institutions including Université de Lorraine and Zhejiang University.
RF-GPT was trained on approximately 625,000 simulated radio signal samples, enabling it to develop a deep understanding of wireless environments.
Toward AI-native 6G
Experts say the model represents a shift from fragmented, task-specific RF tools to a unified system that merges language AI with spectrum intelligence.
According to Debbah, the innovation opens the door to “AI-native radio systems,” where networks can interpret and respond to wireless conditions in real time—an essential step toward future 6G connectivity.
Strategic impact
University officials say the launch reinforces the UAE’s ambition to become a global leader in artificial intelligence and digital infrastructure.
With RF-GPT, telecom operators, engineers, and regulators could soon have a powerful new tool to optimise networks, improve efficiency, and enhance decision-making in an increasingly connected world.
