Paramedics work one of medicine's most overloaded jobs: delivering care, tracking vital signs and hunting for a hospital able to accept the patient, then often reconstructing their activity records from memory afterwards. A Severance Hospital team led by Professor Chang Hyuk-jae of the Severance Cardiovascular Hospital has built an integrated AI platform to carry that load, developed as part of a National Fire Agency R&D project and announced in December 2025.
The platform combines ten AI models across four functional categories. Voice recognition converts on-site dialogue into digital records automatically. Predictive models estimate the risk of patient deterioration during transport. Severity assessment runs on ambulance CCTV footage, producing a preliminary triage evaluation, and field support tools offer treatment guidance plus recommendations for the optimal receiving hospital, with field photos and evaluation notes transmitted ahead of arrival.
In its first phase, paramedics testing the prototype rated ease of use, work efficiency and response speed highly, with overall satisfaction scoring 86 points against an 80-point target, and the hospital recommendation feature singled out as especially useful. The second phase will measure response-time gains and reductions in recording burden quantitatively in real-world settings. The stated end goal is blunt and worthy: getting accurate field information to emergency physicians fast enough to improve survival.