A team of students from Middlesex University Dubai has developed an AI system that lets hospitals collaborate on patient data without exposing sensitive personal records, taking top honours at the Dubai Electronic Security Centre's cybersecurity competition held on the sidelines of GISEC Global 2026. The event ran from September 16 to 18, 2026, at the Dubai Exhibition Centre in Expo City Dubai.
The competition's theme was "Privacy-Preserving Analytics on Sensitive Government Data," and drew more than 400 students across 350 teams in its second stage, with only seven teams advancing to the final. The winning team — Aiman Waheed, Mary Camille Bognot and Nikoloz Khorkheli, all third-year BSc Cybersecurity and Digital Forensics students — presented a project titled "The Model Is The Leak." the system is built around two simulated hospitals, each able to train its own AI model using only its internal data, without transferring the underlying patient records to the other institution. Rather than exchanging sensitive medical information directly, the two hospitals share only the trained model itself.
The students then tested whether patients whose data had trained the model could still be identified from it. Their findings showed the privacy safeguard could be implemented while costing less than one per cent in overall accuracy — a result the team says demonstrates that patient privacy and system performance do not have to come at each other's expense.