Researchers at the University of Wisconsin School of Medicine and Public Health are building an AI-powered “digital mentor” that coaches doctors through newborn intubation, one of the highest-risk procedures in the neonatal intensive care unit.
Preemies frequently require a breathing tube in the windpipe because their lungs haven’t finished developing. It’s a delicate procedure with little room for error. Repeated or drawn-out attempts raise the odds of low oxygen, a slow heart rate, and damage to the airway. Mastery can demand as many as 100 supervised tries, and hands-on chances are scarce, according to Ryan McAdams, a UW pediatrics professor.
The team’s idea is to give trainees an AI coach that can evaluate technique and provide feedback outside the clinical setting, supplementing the rare chances to practice on real patients. Video laryngoscopy, which shows the airway on a display, is spreading and hands the AI a natural view of what the clinician sees.
McAdams said the team recognized a clinical problem: teaching the skill as safely and quickly as possible.
The project is part of a broader push to use AI in medical education, where simulation and feedback loops can stretch limited training time. For NICUs facing staffing pressures, a tool that compresses the learning curve for a high-stakes skill could translate directly into safer care for the smallest patients.
