The American Gastroenterological Association (AGA) has launched Nigel, a point-of-care AI assistant built exclusively for gastroenterology and hepatology. Unlike general-purpose AI tools, Nigel is grounded in AGA's peer-reviewed clinical guidelines, structured around the GRADE evidence framework, and vetted by GI and liver experts — designed to be a trustworthy "AI colleague" rather than just another chatbot.

AI is already reshaping GI practice in multiple ways: Medtronic's GI Genius module detects colorectal polyps in real time during colonoscopy, general LLMs like OpenEvidence and DoxGPT are being used as rapid clinical reference tools, and ambient AI scribes are being piloted to reduce documentation burden. Some institutions have already made AI-assisted colonoscopy the standard of care.

But prominent gastroenterologists from Yale, Columbia, and Baylor College of Medicine are sounding a measured alarm, highlighting risks including inaccurate or outdated outputs, automation bias, potential deskilling of trainees, and patient privacy vulnerabilities.

Key Takeaways:

• AI should augment, not replace clinical judgment — the physician remains responsible for final decisions
• CADe tools like GI Genius improve polyp detection but still require the endoscopist to characterize and manage lesions
• Independent validation and post-implementation monitoring are essential as the field evolves rapidly
• Clinicians must verify AI-cited evidence for accuracy and contextual appropriateness

Why it matters: As AI tools move toward becoming standard practice in GI, getting adoption right — with proper oversight, transparency, and training — is critical for both patient safety and the long-term preservation of clinical expertise.