The integration of artificial intelligence (AI), deep learning (DL), and radiomics is rapidly reshaping gastroenterology and hepatology. Advanced computational models including convolutional neural networks, recurrent neural networks, transformers, artificial neural networks, and support vector machines are revolutionizing both clinical practice and biomedical research. This review explores the broad applications of AI in managing patient data, developing disease-specific algorithms, and performing literature mining. In drug discovery, AI-driven computational chemistry is significantly speeding up drug discovery and development by accelerating hit identification, lead optimization, and formulation development. Machine learning models enable the precise prediction of molecular interactions and drug-target binding, thereby improving screening efficiency and reducing reliance on conventional experimental methods. AI also plays a central role in structure-based drug design, molecular docking, and absorption, distribution, metabolism, excretion, and toxicity simulations, while facilitating excipient selection and optimizing formulation stability and bioavailability. In clinical endoscopy, DL-enhanced computer vision is advancing ambient intelligence by enabling real-time image interpretation and procedural guidance. AI-based predictive analytics further support personalized medicine by forecasting treatment response in inflammatory bowel disease. Remote monitoring systems powered by AI are proving vital in managing high-risk populations, including patients with acute-on-chronic liver failure, liver transplant recipients, and individuals with cirrhosis requiring individualized diuretic titration. Despite its promise, AI potential in gastroenterology faces challenges stemming from data inconsistencies, ethical concerns, algorithmic biases, and data privacy issues including health insurance portability and accountability act and general data protection regulation compliance. Establishing standardized protocols for data collection, labeling, and sharing, alongside robust multicenter databases and regulatory oversight, are essential for ensuring safe, ethical, and effective AI integration into clinical workflows.
Yatin et al. (Thu,) studied this question.