The purpose of this study is to compare the efficiency of writing endoscopic reports by artificial intelligence and physicians through a randomized controlled trial.

Digestive endoscopy permits early detection of gastrointestinal diseases and improving the patients' quality of life. However, writing endoscopic report takes a lot of manpower and time, reducing the efficiency of endoscopy. Our center has developed a structured diagnostic report generation system for digestive endoscopy, which can generates the report by artificial intelligence. Here, we compare the efficiency of writing endoscopic reports by artificial intelligence and physicians through a randomized controlled trial.

Inclusion Criteria:

• patients aged between 18 and 80 years；
• patients undergoing gastroscopy and colonoscopy examination with anesthesia；
• patients able to give informed consent

Exclusion Criteria:

• patients with the contraindications to endoscopic examination;
• a history of gastrointestinal cancers;
• allergic to anaesthetics in previous medical history;
• the endoscopic procedure cannot be completed due to stenosis, obstruction, solid food or complications of anaesthesia.