AI for GI

This month’s Gastroenterology issue is devoted solely to the use/expected uses as well as risks of artificial intelligence (AI) for gastroenterology and hepatology.

DL Shung, M Iacucci. Gastroenterol 2025; 169: 391-392. Artificial Intelligence in Gastroenterology and Hepatology: Potential and Perils

An excerpt:

“AI is reshaping the landscape of gastroenterology and hepatology with the promise of better, faster, more objective, and standardized care of delivery. However, behind the algorithms lies a more insidious risk: the erosion of trust in human providers…Information risk …include both error commission (ie, when the models generate false statements, introduce nonsensical concepts, or fabricate sources) and error omission (ie, summaries that omit critical information)…

When AI becomes the center of care, patients may perceive their doctors as intermediaries…diminishing the therapeutic effect of the patient-physician relationship…This arrangement can dilute clinical training, increase physician burnout, and lead to medicolegal implications…Other risks include perpetuating bias from nonrepresentative training data and amplifying uncertainty of AI due to lack of real-world validation…

We hope that AI systems will allow us to spend more, not less, time with patients and empower us to provide personalized care by leveraging high-quality multimodal data.”

Most of the articles are behind a paywall in this issue. There are five that are open access articles:

My take: These articles provide a good deal of information about the applications and risks of AI. In my view, physicians will be needed more than ever to help interpret/manage the huge amount of information available.

Related blog posts: