Future utility of generative artificial intelligence in colorectal surgery

Abstract

Generative artificial intelligence (GAI), which learns to produce new and original content from large datasets, has been integrated widely into everyday applications. Colorectal surgery represents a significant opportunity for GAI given the large and heterogeneous datasets, such as radiology, histopathology, omics, surgical video and electronic health records, relevant to its practice and training. GAI-powered chatbots, such as ChatGPT, might assist patients and surgeons in the management of malignant and benign colorectal disease in a range of clinical scenarios and pathways. The development of colorectal foundation models has the potential to transform surgical data surplus into clinically meaningful tools with a wide range of applications from intraoperative assistance to simulated training. Other clinical applications include automation of clinical documentation and use of virtual assistants. GAI platforms have also been demonstrated to encode clinical knowledge and have been utilised to create training simulations. Use of GAI within colorectal surgery must, nevertheless, be moderated by colorectal surgeons who understand and account for the inherent risk within de novo generated outputs. Wider adoption of GAI tools requires consideration of technical, clinical, ethical and regulatory obstacles.

Ähnliche Arbeiten