Compelling versus polite reasons to have doctors in the loop with AI decision-making in medicine
As part of the 5th Advanced School in Computer Science and Engineering, the 2023 Rabin Lecture will be delivered by Prof. Isaac Kohane.
Wednesday, 18 January at 5-6.30pm
The lecture will be held in the Computer Science auditorium, located in the Rachel and Selim Benin School of Computer Science and Engineering, Rothberg A, 3rd floor.
ABSTRACT
In 2016, Geoffrey Hinton, one of the pioneers of modern neural network technology famously said that we should stop training radiologists and that neural networks would be outperforming radiologists within 5 years. Since Hinton’s original pronouncement, most computer scientists have learned not to make similar predictions, if only not to antagonize their medical colleagues. Nonetheless there was a solid nugget of accurate prediction in Hinton’s pronouncement but also a misunderstanding of the current limitations of AI in medical practice. I will illustrate this misunderstanding through example medical tasks which require humans-in-the-loop. I will review these limitations and propose steps to leap beyond them in the 3-5 year time-frame by elaborating a simple taxonomy of paths through evaluation and acceptance for AI artifacts in medicine: the Trainee, the Trial, and the Torchbearer. In doing so, I will also review what are the largest challenges facing medicine today and the mismatch as to where the current AI systems are directed. I will outline where Hinton’s premature assessment will eventually pan out and where there will be an irreducible advantage to nimble human experts. Finally, I will outline a necessary tectonic shift that will truly place patients at the center of a community of clinical decision-making and services, one for which AI is only a small but important piece.
ABOUT THE SPEAKER
Isaac “Zak” Kohane, MD, PhD, is the inaugural chair of Harvard Medical School’s Department of Biomedical Informatics, whose mission is to develop the methods, tools, and infrastructure required for a new generation of scientists and care providers to move biomedicine rapidly forward by taking advantage of the insight and precision offered by big data. His most urgent question is how to enable doctors to be most effective and enjoy their profession when they enter into a substantial symbiosis with machine intelligence. He is a member of the National Academy of Medicine, the American Society for Clinical Investigation and the American College of Medical Informatics.
