Elena Tessitore, Staff Physician in the Division of Cardiology  and in the Division of Internal Medicine and Rehabilitation at the HUG and Privat-docent in the Department of Medicine at the UNIGE Faculty of Medicine, and Douglas Teodoro, Assistant Professor in the Department of Radiology and Medical Informatics at the UNIGE Faculty of Medicine and leader of the Data Science for Digital Health Group, will co-lead the healthcare pillar of the European project ARGENTIC — a collaboration that embodies the vision of the Geneva Centre for Cardiac and Vascular Research (GCCVR), recently established at the Faculty of Medicine to foster synergies between clinical expertise and data science. Funded under the Horizon Europe programme, the project has secured €7.5 million in funding, within a consortium of 19 institutions across 9 countries.

ARGENTIC aims to develop a sovereign, trustworthy, and energy-efficient European ecosystem for agentic AI — that is, AI capable of acting autonomously — deployable from the cloud all the way to connected devices at the point of care. Its impact will be demonstrated through seven large-scale pilots across sectors as diverse as healthcare, energy, agriculture, and tourism.

The project relies on federated learning, knowledge graphs, and digital twins to enable collaboration between systems without sharing raw data, while guaranteeing explainability, human oversight, and compliance with the AI Act and GDPR. "In practice, this means that several hospitals can collaborate to train an AI model without sharing their patients' data," explains Douglas Teodoro. "Each hospital trains the model locally and shares only the results of that learning process. Knowledge graphs and digital twins enrich this process by enabling systems to reason, anticipate, and continuously improve, while remaining under human control."

Within this framework, the Geneva teams will lead the healthcare pilot: deploying lightweight, sovereign language models directly on bedside monitoring devices. "With ARGENTIC, we will be able to analyse in real time the vital signs of patients hospitalised for cardiac conditions — ECG, heart rate, blood pressure, oxygen saturation," adds Elena Tessitore. "This will allow us to detect early warning signs of arrhythmia or cardiac decompensation, and alert clinicians before the situation deteriorates — all without patient data ever leaving the hospital."