Most clinical AI is built and validated on adults, and children’s hospitals often find it doesn’t transfer. This panel looks at what it takes to build, validate and scale AI that works for children, across four areas:

• why imaging models trained on adults underperform in pediatric patients
• the gap between FDA-cleared AI devices and the pediatric evidence behind them
• the reality of running deterioration and sepsis tools at the bedside in critical care
• whether federated, multi-institution approaches can pool scarce pediatric data to detect disease earlier

Panelists will share real deployment experience and discuss what the field needs to move from isolated successes to systems that work at scale.