AI adoption in gastroenterology is expanding across clinical decision support, colonoscopy, documentation and evidence review, but gastroenterologists interviewed by Medscape emphasize that implementation needs to keep pace with validation and safeguards. The article highlights Nigel, the AGA’s GI-specific AI assistant, which is built around AGA guidelines and the GRADE evidence framework and was reviewed by GI and liver experts.
AI-assisted colonoscopy is another established use case. CADe systems such as GI Genius can highlight potential polyps in real time and have been associated with higher detection of precancerous lesions. However, clinicians raise concerns about deskilling, automation bias, inaccurate or outdated outputs and privacy, particularly as AI becomes more embedded in routine workflows.
The gastroenterologists interviewed stress the need for independent validation, post-implementation monitoring and clinician oversight. General-purpose AI can rapidly synthesize literature and generate differential diagnoses, but outputs may be incomplete, overly confident or poorly suited to an individual patient’s circumstances. Ambient AI scribes may reduce documentation burden, but generated notes still require physician review before entering the medical record.