Experts at the 2026 ESPGHAN Congress emphasized that AI in pediatric gastroenterology should remain a decision-support tool, not a substitute for physician judgment. The most promising applications include AI-assisted endoscopy, histology, imaging, and monitoring for IBD relapse, as well as nutrition-related tasks such as malnutrition screening and parenteral nutrition planning.
A key caveat is that adult AI models cannot simply be transferred to children. Pediatric-specific validation, prospective trials, harmonized datasets, and ongoing monitoring for model performance and drift are needed before wider clinical adoption. In steatotic liver disease, the discussion also highlighted that simpler models using routine clinical and laboratory data can sometimes perform as well as or better than models built on complex genomic or metabolomic data.