Abstract
Artificial intelligence (AI) is reshaping modern medicine, and gastroenterology and hepatology are among the specialties where its impact is becoming increasingly evident. AI has demonstrated the ability to process and analyze large amounts of clinical, radiological, endoscopic, and multi-omics data, offering unprecedented opportunities to enhance diagnostic accuracy, optimize therapeutic decision-making, and reduce variability in clinical practice. In endoscopy, computer-aided detection and diagnosis systems have shown consistent improvements in adenoma detection rates and real-time polyp characterization, while in hepatology, machine learning models outperform traditional scores for non-invasive assessment of liver fibrosis. Furthermore, multimodal approaches integrating genomics, microbiome, and imaging data are paving the way for precision medicine in inflammatory bowel disease and other complex digestive conditions. Despite these promising advances, significant barriers remain. The quality and heterogeneity of training data, the lack of rigorous external validation, and the opaque "black box" nature of many algorithms limit their clinical reliability. Ethical challenges, including accountability in case of diagnostic errors, protection of patient privacy, cost, and equitable access, also need to be addressed. This narrative review summarizes the current applications of AI in gastroenterology and hepatology, critically examines methodological and ethical challenges, and outlines future perspectives. Responsible, transparent, and equitable implementation will be essential for AI to transition from an emerging promise to a consolidated tool that improves outcomes and advances personalized digestive care.