AI project seeks major leap in diabetes care

Clinical trials across leading US research centres will test adaptive systems that learn from real-time data to personalise diabetes care.
Clinical trials across leading US research centres will test adaptive systems that learn continuously from real-time patient data.

A major research initiative led by the University of Virginia has secured $4.7 million to advance machine learning in Type 1 diabetes care.

The project, backed by Breakthrough T1D and the Helmsley Charitable Trust, aims to develop fully automated insulin systems that adapt continuously to patient needs.

The research will combine adaptive algorithms with ultra-rapid insulin to enable personalised glucose control without manual input. The University of Virginia will lead engineering and algorithm development, with clinical trials conducted across multiple US research centres.

At its core is an AI framework that learns from real-time data, adapting to metabolic changes, stress, and daily rhythms. Researchers aim to overcome the limitations of current automated insulin systems, which still rely on fixed parameters and regular user intervention.

The collaboration reflects a shift towards patient-centred AI, aiming to reduce daily diabetes management burdens while improving safety and quality of life. Developers say the technology could offer families greater freedom and long-term stability in managing chronic conditions.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!