STAT+: Why this former ER doc thinks giving AI less information is sometimes better

Brittany Trang, Ph.D., covers AI in health and medicine: Does it actually work? Who benefits, or might be harmed? She writes the weekly AI Prognosis newsletter. Follow her on Threads, Mastodon, and Bluesky. You can reach Brittany on Signal at btrang.01.Let me know if, especially after reading the below item on what recent heart drug failures mean for AI in drug discovery, you have questions about AI in pharma. I’ll be interviewing executives from Novartis and Johnson & Johnson on stage next week at the HLTH conference. And if you have questions for Epic, come on down to the STAT stage takeover at HLTH a week from today, where I’ll be talking to Aneesh Chopra, Brendan Keeler, and Epic executives Antonia Papajani and Seth Howard. Less is more? In today’s clinical AI landscape, every platform is making a land grab: Epic, traditionally a system of record, now wants to display clinical recommendations. Abridge, which started as an ambient scribe, wants your patient’s insurance information. OpenEvidence, which originally answered clinical questions, wants you to enter patient records and visits so it can give you personalized recommendations. STAT+ Exclusive Story Already have an account? Log in This article is exclusive to STAT+ subscribers Unlock this article — plus in-depth analysis, newsletters, premium events, and news alerts. Already have an account? Log in View All Plans To read the rest of this story subscribe to STAT+. Subscribe