Artificial intelligence (AI) has demonstrated a huge potential to transform the real-world practice of gastrointestinal endoscopy: i) improved quality metrics and reduced lesion miss rate in AI-assisted colonoscopy, ii) guided anatomical structure recognition and lesion segmentation in hepatobiliary endoscopy, iii) reduced reading time and enhanced lesion detection in capsule endoscopy, iv) critical structure identification in advanced therapeutic endoscopy. It will enhance training efficiency, standardize outcomes, and ultimately improve patient care.
From intra-procedural guidance to post-procedural feedback, AI can address the current challenges that we are facing in education. However, the major concerns before a universal implementation include risks of deskilling, over-reliance and ethical considerations. Governing frameworks must address data privacy and transparency, algorithmic bias and interpretability, and legal accountability.
In the future, the next generation endoscopy training requires a stepwise approach to integrate AI into traditional mentorship, aiming to balance technical and humanistic competencies.
