Berzin, a senior endoscopist at BIDMC and an associate professor of medicine at Harvard Medical School, has led several early national and international studies exploring the role of AI in precancerous colorectal polyp detection. This is a “level 1” assistance algorithm. But AI tools may soon go beyond polyp detection to help predict polyp diagnosis, potentially eliminating the need for tissue biopsy. The authors suggest that even more advanced levels of automation are technically feasible and imminent, potentially providing automated endoscopic reports and recommendations to assisting physicians.
Lead author Sami El-Amin, a clinical research fellow in the department of gastroenterology at BIDMC and Harvard Medical School, used hypothetical scenarios to explore the potential legal liability of individual physicians or healthcare institutions for a variety of AI-generated errors that could occur in the field of gastrointestinal endoscopy.
The authors conclude that the degree of legal liability for AI errors will depend on how these tools are integrated into clinical practice and the level of automation of the algorithms. Collaboration between hospitals, medical groups, and gastroenterologists is essential to ensure the safety, proper implementation, and oversight of these AI tools. Specialty societies and health care organizations should establish guidelines for physician oversight of AI tools at various levels of automation. For physicians, whether they are following or deviating from AI recommendations, meticulous clinical documentation remains the cornerstone to minimize liability risk.
sauce: Beth Israel Deaconess Medical Center
