AI and Clinical Practice-AI and the Ethics of Developing and Deploying Clinical AI Models

AI in clinical practice needs ethical frameworks to avert future biases. In this Q&A, Marzyeh Ghassemi, PhD, the Herman L. F. von Helmholtz Career Development Professor at MIT in Electrical Engineering and Computer Science (EECS), joins JAMA's Editor in Chief Kirsten Bibbins-Domingo, PhD, MD, MAS, to discuss ethical machine learning and responsible clinical implementation.
Click ja.ma/3uro4nN to read AI Developers Should Understand the Risks of Deploying Their Clinical Tools, MIT Expert Says in JAMA.
Topics discussed in this interview:
00:00 Introductions
01:15 Biases and Artificial Intelligence
02:24 Why a model might not perform?
06:52 Labels are important for how AI models are trained.
08:24 Ethical Machine Learning
10:26 The Risks Associated With How An AI Model Is Deployed
13:43 Models Deployed Ahead of Regulation
18:23 Automation Bias
19:39 Explaining the model can give clinicians better insights.
22:22 AI Modeling vs Clinical Risk Scores
23:58 What AI tools do you use?
24:28 Conclusion
---------------------------------------------------------------------------------
For more from JAMA
• www.jama.com
• / jamajournal
• / jama_current
• / jama. .
Follow the JAMA Network
• www.jamanetwork.com
• www.jamanetworkaudio.com
• / jamanetwork
• / jamanetwork
• / jamanetwork
• / jamanetwork
#JAMAai

[@AndreaGarcia-zl1hn]

Great interview! Very insightful! Thank you.

[@angie.veitch]

My #1 (from minority health) is: Language.