Gastroenterology and Artificial Intelligence: 3rd Annual Artificial Intelligence Summit (On-Demand) | September 2021-How to Evaluate AI in the GI Literature and Clinical Trials

Video Transcription

Video Summary

The video transcript is a presentation on the topic of artificial intelligence (AI) in gastrointestinal (GI) endoscopy. The speaker discusses the different phases of the innovation cycle for AI in GI endoscopy and the importance of evaluating its performance in real-life settings. The speaker emphasizes the need to consider various factors when evaluating AI systems, such as the training data set (including patient population, centers involved, and disease verification), the testing data set (including generalizability to community-based endoscopy and comparison with non-expert endoscopists), and the potential biases that can affect the results. The speaker also discusses the need for randomized trials to assess the effectiveness of AI in real-world situations and highlights the importance of measuring the value of AI innovation in terms of costs, cancer prevention, and increased standard of care. The presentation concludes with a reminder about the importance of considering the details of training and testing databases, the interaction between machines and humans, and the need for value-based assessment of AI in GI endoscopy. The speaker expresses gratitude for the invitation and welcomes further discussion on the topic. No credits are specified in the video.

Asset Subtitle

Cesare Hassan, MD

Keywords

artificial intelligence

gastrointestinal endoscopy

evaluation

biases

effectiveness