AI’s Bias in Predictive Healthcare with Associate Professor Marzyeh Ghassemi

episode

Building 32 from MIT CSAIL Alliances not yet transcribed

▲ 0

Transcript

This episode hasn't been transcribed yet Upvotes decide what gets transcribed next — 0 so far. Sign in to upvote

Related

recent transcribed audio

Something is seriously wrong in the Northern Territory 18 min · 13 Sep 2026 · 7am

Raygun and Glory Holes 52 min · 13 Sep 2026 · You Wouldn't

The Saint: The Strange Bedfellows 04/15/1951 30 min · 13 Sep 2026 · Choice Classic Radio Detectives | Old Time Radio

Johnathan Thurston: The Rise of One of the NRL’s Greatest Halfbacks & Life After Footy 1h 11m · 13 Sep 2026 · Straight Talk with Mark Bouris

Electric Cars: Connected Car Security fears and plans 38 min · 13 Sep 2026 · Two Blokes Talking Tech

Ep. 581 - Bushpig Campfire Whiskey feat. Danny Grant 2h 47m · 13 Sep 2026 · Alpha Blokes Podcast

Report any issue

Source

Recorded27 Nov 2024

Publisherfeeds.soundcloud.com ↗

Handlerec:2204985

Show notes

Associate Professor Marzyeh Ghassemi discusses why, despite its tremendous potential for good, AI must be approached with caution, especially in high risk areas like healthcare. Biases in training data can perpetuate real-world inequalities, providers might rely too much on potentially faulty AI, and tools created for one demographic might not necessarily translate to a different population. However, considering the incredible potential of AI to tackle challenges in transformative ways, it’s worth approaching these considerations with rigorous academic insight. Learn more about her work here: https://www.csail.mit.edu/person/marzyeh-ghassemi

Ask this episode

“What did the council decide on item 42?” — answered with the speaker and the moment.

Sign in